Abu Dhabi’s 1 GW AI Campus Needs More Power Than the Country Forecast SeaPRwire

Abu Dhabi’s 1 GW AI Campus Needs More Power Than the Country Forecast

Inside the 200 MW first phase, the 6 TWh national forecast and the gas, nuclear and solar capacity Abu Dhabi is lining up for 2026 to 2030 The first 200 MW of Stargate UAE are supposed to be live this year. The site sits inside a 26 square kilometre campus in Abu Dhabi, built by G42 through its subsidiary Khazna, operated with OpenAI and Oracle, and fitted with roughly 100,000 Nvidia Grace Blackwell GB300 chips. The partners put the investment at up to $10 billion for a 1 GW cluster. That cluster is the easy part to describe. The hard part is the meter. A megawatt tells you how fast a facility draws power. It does not tell you how much electricity it will consume in a year, and it does not tell you whether the grid can deliver that power at three in the morning in August, when the air conditioning load across the emirate peaks and the solar farms are dark. Those two questions decide whether the promised capacity arrives on schedule, and how much new generation Abu Dhabi has to build around it. The public numbers on both sides are already in circulation. They are rarely put in the same sentence. The 200 megawatt number hides the annual bill Start with the chips, because they set the floor. TrendForce estimated that a GB300 system with 72 chips draws about 140 kilowatts. At that rate, the 100,000 chips announced in May 2025 work out to roughly 1,400 systems and about 196 MW of information technology load, which is why the partners landed on 200 MW for the first phase in 2026. Convert the IT load to what the utility sees. Modern desert builds in the UAE are designed at a power usage effectiveness of about 1.18, so a 2026 design of 200 MW of chips becomes roughly 236 MW at the fence line. Run that for a full year and the phase consumes about 2.1 TWh. Two point one terawatt hours is not a rounding error in this market. Wood Mackenzie estimated that every data centre in the UAE consumed about 3 TWh in 2025, roughly 2% of national electricity demand of 173 TWh. The first phase of one project, running flat out, would add two thirds of the entire country’s existing data centre consumption. Scale it up and the arithmetic gets louder. The full 1 GW Stargate cluster becomes 1.18 GW at the fence and about 10.3 TWh a year. The 5 GW campus announced in 2025 becomes 5.9 GW and about 51.7 TWh a year, close to 30% of the country’s total 2025 demand. These are ceilings, not forecasts. Training clusters do not run at 100% load every hour, and no operator signs up for that. Cut the utilisation to 70% and the 1 GW cluster still needs about 7.2 TWh a year. Cut it to 50% and it needs 5.2 TWh, which is more than the 6 TWh that the whole national data centre estate is projected to consume in 2030. Table 1: Load converted to annual energy at an assumed design PUE of 1.18 and full utilisation Load IT capacity (MW) Facility load (MW) Annual energy (TWh) Share of 2025 national demand Stargate phase 1 200 236 2.1 1.2% Full Stargate cluster 1,000 1,180 10.3 6.0% UAE-US AI Campus 5,000 5,900 51.7 29.9% UAE data centres, 2025 actual n/a n/a 3.0 1.7% Source: Wood Mackenzie via Data Center World Middle East, with conversions calculated from published capacity figures The last row is the one to keep in view. The country’s entire data centre fleet used 3 TWh last year. Its flagship AI project is designed to use several times that on its own. Two forecasts, one grid In March 2026, Wood Mackenzie put UAE data centre demand at just over 6 TWh by 2030, more than double the 2025 level. EWEC, the Abu Dhabi utility that procures the emirate’s power and water, has used a similar figure in its own planning work. Read that forecast next to the campus arithmetic and the two numbers do not fit. The 6 TWh estimate was built on colocation demand: the racks that banks, telcos and cloud regions rent today, plus growth in the same shape. A 5 GW AI campus is not a bigger version of that business. It is a single customer whose load arrives in 200 MW blocks and whose utilisation curve is set by GPU replacement cycles, not by office hours. The difference matters for anyone selling into this market. If you plan against 6 TWh, you size transformers, chillers and backup generation for a country that adds roughly 3 TWh of new data centre load over five years. If you plan against the campus pipeline, you are sizing for a single site that could pass that number before 2030. Both numbers can be right at the same time. The forecast describes the market that exists. The campus describes the demand that has been announced but not yet energised. The gap between them is the thing nobody has published: how much of Stargate’s 1 GW is already inside the 2030 figure of 6 TWh, and how much sits on top of it. My reading is that the 6 TWh is a colocation forecast with an AI campus stapled to it, and that the first revision will come from the utility rather than the researchers. Larry Ellison, Oracle’s chief technology officer, described the intent at launch: “This first-in-the-world platform will enable every UAE government agency and commercial institution to connect their data to the world’s most advanced AI models.” That is a statement about demand, not about supply. There is a clue in the way the project is being built. Khazna, the G42 data centre arm that holds about 60% of the UAE’s commercial wholesale capacity, is the builder. Its own pipeline target is 1,000 MW. Stargate’s first phase, due in 2026, would absorb a fifth of that. When the same company is both the landlord for the national forecast and the contractor for the campus, the two demand curves get blended in a boardroom long before they reach a public statistic. Firm power is the scarce item, not sunshine Abu Dhabi has an unusual starting position for a Gulf AI hub. It is the only country in the Middle East with an operating nuclear fleet. Barakah supplies about 21% of UAE electricity output, and EWEC lists 4,170 MW of Barakah capacity in service with another 1,390 MW to come. Solar contributed about 9% of generation in 2025, with official projections taking that above 20% by 2040. Neither source solves the problem on its own. Nuclear runs flat and cannot follow a training cluster that ramps when a batch job starts. Solar runs for the middle of the day and stops. The campus needs both energy and firm capacity, and the firm capacity is what costs money. The procurement list shows how the emirate is filling that gap: Taweelah C, a 2.6 GW combined cycle gas plant at the Al Taweelah complex, with Siemens Energy supplying three SGT5-9000HL gas turbines and two steam turbines. TAQA leads it with Al Jomaih and Sembcorp, China Energy Engineering is the EPC contractor, and commissioning is expected in late 2028. The round the clock renewable project developed by Masdar with EWEC, which pairs 5.2 GW of solar with 19 GWh of battery storage to deliver 1 GW of firm output. Financial close came in July 2026 at about $6.1 billion, with operations targeted for 2027. Barakah’s remaining units, which add baseload without new fuel logistics. Put the gas plant next to the campus and the scale becomes concrete. A 2.6 GW combined cycle plant running at a 60% capacity factor from 2029 produces about 13.7 TWh a year, which is more than the full 1 GW Stargate cluster needs even at full load. One power station, built for grid flexibility and carbon capture readiness, covers the flagship cluster with room to spare. The battery project is the more interesting bet, because the arithmetic is tight. Nineteen gigawatt hours of storage against 1 GW of promised firm output is 19 hours of delivery at full rate. That is enough to carry a night. It is not enough to carry a week of heavy dust or a gas outage, and it explains why the gas plant is still in the plan. What the campus buys before it buys chips The conventional way to read a project like this is as a chip order. The operational way is as a procurement queue, and the first items on it are not silicon. A 200 MW phase needs high voltage substations, transformers, switchgear and chillers before a single rack is installed. Those items have lead times measured in years, not months, and they compete with every other Gulf megaproject for the same factory slots. G42 said in October 2025 that civil, structural and mechanical work was ahead of schedule and that all long lead equipment had been procured. That sentence is the most informative line in the whole announcement, because it tells you the constraint the company was managing. Cooling is the second queue. A design PUE of 1.18 in a desert climate implies liquid cooling at the rack and closed loop heat rejection, not evaporative cooling that drinks water. That changes the equipment list: coolant distribution units, rear door heat exchangers, dry coolers sized for 45 degree ambient air. It also changes the maintenance contract, because a coolant leak in a 140 kW rack is a fire risk, not an inconvenience. The third queue is people. A campus of this size needs mechanical and electrical technicians, network engineers and security staff on site around the clock, in a country where the data centre workforce is thin. The regional buildout is adding capacity faster than it is adding trained staff, and that gap shows up as higher operating costs and longer commissioning. For suppliers, the sequence matters more than the headline number. In the 2026 phase, the first $1 billion goes to switchgear, transformers, cooling and construction. Chips are the last thing to arrive and the first thing to be quoted. The licence that came before the concrete Stargate UAE exists because Washington changed a rule. The Biden administration had restricted advanced AI chip exports to countries including the UAE. The Trump administration rescinded that rule in 2025, and the Commerce Department set up a bilateral working group to check that the project met US security standards. That sequence explains the shape of the deal. The UAE side committed to reciprocal investment in US AI infrastructure, and the US side opened the export channel. The campus is the visible half of a bargain whose other half sits in American data centres. The licence also sets the pace of the chip pipeline. Microsoft’s experience is the clearest example: a September 2025 export licence authorised 60,400 A100-equivalent units, on top of the 21,500 it had already deployed. Every one of those units has to be ordered, shipped, racked and cooled before it earns anything. For anyone modelling the project, the lesson is that the constraint moves. In 2025 it was the export licence. In 2026 it is long lead electrical equipment and grid connection. In 2028 it may be the gas turbine queue that is already stretching into the early 2030s in other markets. Khazna is the bottleneck and the beneficiary Khazna’s position explains a lot about how this market actually works. It builds the campus, it holds the majority of wholesale colocation capacity, and it is the delivery vehicle for Microsoft’s 200 MW expansion. When one company is the landlord, the contractor and the operator’s local partner, capacity gets allocated by relationship as much as by price. The Microsoft numbers show the pattern. The company’s cumulative UAE commitment has reached $15.2 billion, including $7.9 billion for 2026 to 2029, of which $5.5 billion is AI and cloud infrastructure. A separate $544 million hyperscale site in Dubai is being built with du, and AWS has committed $1 billion with e&. Add Stargate’s 1 GW and the colocation market forecast starts to look conservative. RAX puts the UAE colocation market at $1.39 billion in 2026, rising to $3.90 billion by 2030 at a 21.7% compound annual growth rate. That forecast was built on leased racks sold by the megawatt. It does not model a single tenant taking 1 GW of dedicated capacity, because until this year there was no such tenant. The other side of the same coin is pricing power. With 400 MW of colocation capacity recorded in 2026 and a pipeline target of 1 GW, Khazna can price scarcity. Tenants who need capacity in 2027 are not negotiating against a competitive market. They are negotiating against a queue. Connectivity softens the picture a little. The UAE hosts 19 international submarine cables, with fibre to the home at 99.5% coverage in 2024. The country is a genuine transit point between Europe, Asia and Africa, which is why hyperscalers keep landing here rather than in a cheaper neighbour. The free subscription that makes curtailment expensive One detail in the Stargate package sits outside the engineering and inside the politics. The UAE became the first country to offer ChatGPT to its residents nationwide, and the project carries a commitment to provide ChatGPT Plus to citizens and residents at no charge. The service lists at $20 a month. That commitment changes the cost of failure. A delayed commercial data centre annoys its tenants. A delayed national AI platform that every resident has been promised annoys everyone, and it does so in public. The operators have an incentive to energise whatever capacity exists, even at poor utilisation, rather than wait for the optimum. It also gives the project a social return that is easy to defend in budget negotiations. The electricity bill for a 1 GW campus is large, but it is being spent inside a country that treats AI capacity as public infrastructure, the same way it treats desalination and the nuclear fleet. That framing is why the gas plant, the battery project and the campus are moving on parallel tracks rather than in sequence. The risk sits in the sequencing. Taweelah C is not expected to commission until late 2028. The round the clock renewable project targets 2027. The first 200 MW is meant to be live this year. For at least one summer, the campus will be drawing on a grid whose new firm capacity has not arrived yet, backed by a nuclear fleet that is already committed and a solar fleet that stops at sunset. Three numbers to watch before 2028 The project was announced with one test attached: if no matching procurement lands before 2028, construction is running behind the story. That test is now measurable. Watch three things. Substation and transformer awards tied to the Abu Dhabi campus, not to the Dubai or Microsoft projects. The commissioning date of Taweelah C, because the gas plant is the firm capacity behind the 1 GW cluster. Any revision to the 6 TWh national forecast, which is the only public number that would have to move if the campus is counted on top of the colocation market. Until one of those moves, the plain reading is that Abu Dhabi has announced more AI load than its published demand forecast contains, and has arranged the generation to serve it on a schedule that runs one to two years behind the buildings. That is not a failure. It is what building infrastructure at this speed looks like. Data basis Data as of November 2026. Capacity figures come from G42, OpenAI and Oracle announcements as reported by Reuters and Khaleej Times, and from the project description published by Data Center World Middle East. National demand of 173 TWh and the 3 TWh to 6 TWh data centre range come from Wood Mackenzie’s March 2026 analysis. Energy conversions multiply facility load by 8,760 hours and assume a constant draw, which overstates consumption for any cluster that idles; the utilisation cases above show the range, not a prediction. The PUE of 1.18 is a design average for modern UAE builds, not a metered figure for this site. Stargate capacity and investment are targets disclosed by the partners, not completed assets. Sources Data Center World Middle East , the strategic case for UAE data centre investment, 2026 Khaleej Times , Stargate UAE to begin in 2026, May 2025 EWEC , generation portfolio and Barakah capacity listing, accessed November 2026 (site home) Enerdata , TAQA selects Siemens Energy turbines for Taweelah C, June 2026 RAX , AI data centre demand in the UAE, 2026 About Global Infinity Labs Global Infinity Labs is an AI innovation company specializing in MCP data services. This article was prepared by our expert team using insights and data provided through our MCP platforms. Contact: support@globalinfinitylabs.com
More
The Wallet Is Live. The Chargeback Rules Are Not. SeaPRwire

The Wallet Is Live. The Chargeback Rules Are Not.

OpenAI’s ChatGPT Wallet shipped on 13 November 2026 with Visa, Mastercard and Amex tokens; the dispute evidence merchants need for agent orders does not exist yet On Friday 13 November, OpenAI turned on ChatGPT Wallet. A US subscriber on the Pro plan, 18 or older, can now add a card issued by Visa, Mastercard or American Express, ask one of OpenAI’s shopping agents to find something, and approve the purchase with Face ID or a fingerprint. The card number never reaches the agent. The network issues a token scoped to that transaction, and the merchant receives a shipping address and a confirmation. The launch is small on purpose. It covers US Pro subscribers only, on US-issued cards, and OpenAI says it is not taking a cut of transactions to start. Chaya Nayak, who leads product for monetization and payments at OpenAI, told CNBC: “Our goal here is to really focus on merchants and our network partners and start building out a really great foundation that’s safe and easy to use.” Everything in that paragraph works. The part that does not work yet is the argument that happens three weeks later, when a customer looks at a statement, does not remember ordering a $340 blender, and calls the issuer. A token proves who paid, not what was agreed Payment tokens are a well understood piece of plumbing. Visa and Mastercard have issued them for years to keep card numbers out of merchant databases. A token scoped to one merchant, one amount and one agent is a modest extension of the same idea, and the networks have built exactly that: Mastercard’s agentic tokens, Stripe’s shared payment tokens, Amex’s agent programme. The token answers a narrow question. It proves that a card was presented by an agent that the issuer had registered, acting for a cardholder who authenticated with a passkey or a biometric. That is a real improvement over what agent commerce looked like a year ago, when assistants were simply given a password and told to log in. It does not answer the question a dispute actually turns on. Did the agent buy what the cardholder asked for? A human checkout leaves a trail: a browser session, a device fingerprint, a cart the person filled, a delivery address the person typed. An agent checkout leaves a token, a mandate and a model. The signals that a merchant normally uses to defend a chargeback are the exact signals that the agent exists to skip. That is not a technology gap that another release closes. It is a gap in the record of intent. No network has a code for this Ask the card networks what happens when an agent order is disputed, and the honest answer as of November 2026 is that nothing specific happens. There is no dispute reason code for an agent-initiated purchase. Worldpay’s July 2026 analysis found no formal liability shift for agent purchases, which means the transaction follows ordinary card-not-present rules. Visa and Mastercard have both published trust frameworks. Neither has published a rule that moves the loss. What they have published is verification machinery: Visa Intelligent Commerce, live since July 2026 with more than 30 European issuers and merchants including lastminute.com, Frasers and Cleverbridge, built around the Trusted Agent Protocol that lets a merchant recognise a registered agent before it buys. Mastercard Agent Pay, which issues agentic tokens to registered agents and, with Google, announced Verifiable Intent in March 2026 to create a cryptographic record linking the cardholder, the instructions and the transaction. EMVCo’s draft framework for card-based agent payments, published on 1 September 2026, built around what it calls Intent Services: a shared record of what the consumer authorised that persists before, during and after a transaction. Read that list carefully. Every item is about identity and authorisation. None of them is a chargeback rule, and none of them creates a new place to file the loss. The practical consequence is that agent orders land in the same bucket as any card-not-present sale, with less evidence than usual. A merchant’s dispute ratio is calculated across all sales, so a new stream of poorly evidenced transactions shows up in a number that acquirers watch. The volume is small today. The direction is not reassuring. Nobody has published what share of agent orders end in a dispute, which is why every number in this market is a projection built on a sample nobody outside the networks can see. What a merchant actually holds after the sale The evidence playbook for agent orders is being written by the people who process the disputes, and it is specific. A merchant defending an agent chargeback needs the agent platform’s identity for that session, the mandate showing what the cardholder authorised, a comparison of the order against the customer’s stated limits, a timestamp showing the customer was notified and could cancel, and proof of delivery to the customer’s address. Most merchants have none of those fields in their order record. They have a token reference and a shipping address. The workaround that is circulating among payment teams is to treat an agent order as a delayed-capture transaction. Authorise the card, hold the capture for 15 to 60 minutes, and push a notification to the human: your agent just ordered this, tap here to void it. A void on an unsettled authorisation costs nothing. A refund after capture costs a fee, and a chargeback costs the fee plus the goods in many cases. That design has a cost of its own. Holding capture means holding inventory, and it slows down the one thing agents are supposed to speed up. It also assumes the cardholder reads the notification, which is the same assumption that failed for every fraud alert of the last decade. None of this makes agent commerce unworkable. It makes it a merchant operations project before it is a consumer habit. The 1.5% that everybody quotes The forecasts for this market disagree by a factor of fifty, and the disagreement is almost entirely about definitions. That is worth knowing before anyone plans against a number. EMARKETER puts US retail ecommerce running through AI platforms at about $20.57 billion in 2026, roughly 1.5% of the total. Morgan Stanley’s base case has agentic shoppers at about 1% of US ecommerce spending in 2026, rising to 10% by 2030, which it values at $190 billion to $385 billion. Bain counts purchases that agents initiate, influence or complete, and lands at $300 billion to $500 billion. McKinsey counts revenue that agents orchestrate, including discovery, and gets $900 billion to $1 trillion in the US and $3 trillion to $5 trillion globally. Table 1: 2030 forecasts for agentic commerce and what each firm counts Firm 2030 US estimate Share of US ecommerce What the number counts Morgan Stanley $190bn to $385bn 10% to 20% Spending by shoppers who use agents Bain $300bn to $500bn 15% to 25% Purchases agents start, sway or finish McKinsey $900bn to $1tn not stated Revenue agents help orchestrate, goods only EMARKETER $20.57bn in 2026 1.5% in 2026 Sales completed through AI platforms Source: Morgan Stanley, Bain and McKinsey forecasts as compiled by Retention Edge, with EMARKETER’s 2026 estimate from the same compilation The spread is not a failure of research. It is what happens when four firms count four different things, and the widest definition wins the headline. A merchant reading “one trillion dollars by 2030” and a merchant reading “1.5% of ecommerce today” are looking at the same market through different lenses. The measured behaviour sits closer to the small number. A Visa survey conducted by Harris Poll in September 2026 found that 72% of US consumers had used an AI assistant, but only 23% trusted generative AI to handle a payment on their behalf. Accenture’s January 2026 survey of 25,590 consumers across 16 countries found 74% would let an agent handle routine tasks and 32% would let it choose what to buy, but only 9% were open to a fully autonomous purchase. People will let an agent do the research. They are not yet ready to let it spend. Where the agent checkout already works The counter-example sits in China, and it is worth studying because it removes the exact variable that is blocking the US market. Alibaba’s Qwen assistant handled more than 120 million orders during six days of the 2026 Lunar New Year shopping period. That volume did not come from a new payment rail. It came from an assistant embedded in a platform that already held the customer’s payment credential, the merchant’s catalogue and the dispute process. When one company owns discovery, checkout and the refund, an agent order has no liability seam to argue about. The platform can reverse a purchase in the app, and the customer never calls an issuer. The cost of that convenience is concentration, which is a different problem from the one OpenAI is solving. Western markets have split the roles. The assistant is OpenAI’s, the credential is a bank’s, the merchant is someone else’s, and the dispute is a network’s. Each party can verify its own step. None of them owns the whole transaction. The conversion data hints at what is being given up. Adobe Analytics found that AI-referred traffic to US retail sites converted 42% better than other channels in March 2026, after converting 38% worse a year earlier. Traffic that arrives with a stated intent is worth more than traffic that arrives from a banner. That is the prize, and it explains why merchants keep testing agents even while their fraud teams object. The adoption numbers inside the assistants are just as lopsided. Meta’s Muse passed 10 million users, with AI-assisted shopping among its most used features, while DoorDash and Uber have reported low numbers of orders placed by agents. People use assistants to decide. They still click the buy button themselves. Two blocks and one withdrawal The resistance is not theoretical. It arrived in the same week as the wallet. On 12 November, United Airlines said it would restrict AI agents from booking flights and managing reservations on a customer’s behalf, citing cybersecurity risk and the federal rule that requires a human to complete check-in and confirm that bags contain no banned substances. The airline said it would block unauthorised third-party agents directly on its platforms and indirectly through travel agencies and aggregators. The week before, Bank of America told customers not to share their passwords with AI assistants. Both moves point at the same hole: an agent that logs in as the customer is indistinguishable from the customer, until something goes wrong. Then there is OpenAI’s own history. The company launched Instant Checkout in the autumn of 2025 and charged merchants 4% of the transaction. By March 2026 it had stepped back from in-chat checkout and moved purchases to the merchant’s own site. Walmart’s Daniel Danker told Wired that conversion was three times lower for items bought inside ChatGPT’s checkout than for items that sent shoppers to Walmart’s own site. The wallet’s zero take rate is the response to that experiment. OpenAI is not monetising the transaction. It is buying the right to be the place where the transaction starts, and paying for it with margin it could have taken. Nayak’s framing, that the focus is merchants and network partners, is the language of a company rebuilding trust it spent once already. Zero take rate is a price, not a gift Work out what the zero costs. If every dollar of the $20.57 billion that EMARKETER expects to flow through AI platforms in 2026 carried a 4% platform fee, the annual revenue at stake would be about $823 million. OpenAI has chosen to forgo that for now, in exchange for merchant relationships and a position in the checkout flow. That is a rational trade for a company that already lost a year to a checkout product merchants did not want. It is also a warning for anyone building on top of the wallet. A platform that is not charging for payments will charge somewhere else, and the likely places are advertising, product placement and the data that comes back from the order. The fee stack explains why merchants are cautious rather than enthusiastic. A 4% platform fee does not replace the card fees a merchant already pays. It sits on top of them. For a retailer running thin margins on physical goods, four points of revenue is most of the profit on the order, which is why the Instant Checkout pilot had a dozen Shopify merchants after five months and not a thousand. The networks have the opposite problem. They are charging for verification, through programmes, certifications and token issuance, and they are not charging for the liability. When an agent order goes wrong, the merchant pays, the issuer fields the call, and the network keeps its fee. The asymmetry explains why the protocols are proliferating. Visa’s Trusted Agent Protocol, Mastercard’s Verifiable Intent, Google’s AP2, Stripe’s shared payment tokens and OpenAI’s own checkout specification all describe the same transaction from different angles. A merchant that wants to accept agent orders in 2027 may need to support four of them. Europe has a rule already, and it points at the human Europe starts from a different place, because it already regulates who must authenticate. The second Payment Services Directive requires strong customer authentication of the payer. When an agent acts for a payer, the regulation still looks through the agent to the person behind it. Visa’s European pilots handled that with payment passkeys bound to the user, which is why Intelligent Commerce could go live with 30 issuers before the US had a consumer product. That framing is useful and incomplete. Strong authentication proves the human approved the transaction. It does not define who is responsible when the agent approves something the human did not intend. The directive was written for third-party payment providers: companies with a licence, capital and a supervisor. An agent has none of those things, and no rule published through 2026 assigns it a legal personality. The gap will be tested first in categories where liability chains are already explicit. Insurance, banking and healthcare all have processes for who pays when a sale goes wrong, and all three are seeing rapid agent adoption. Nearly seven in ten American shoppers already use AI somewhere in the buying process, and about 30% have completed a purchase through it. Those numbers describe a consumer habit that is arriving faster than the rules that govern it. One rule change would matter more than any protocol Three developments would move agent payments from a controlled experiment to a real channel, and each is observable. A network publishes a dispute reason code for agent-initiated purchases, or a liability shift attached to a certified agent programme. That single rule change would do more than any protocol release. A large merchant group adopts delayed capture as standard for agent orders, which would turn the notification step into a consumer habit rather than a fraud control. A regulator states whether an AI agent acting for a consumer is a payment initiation service under PSD2 and its successors. The directive was written for licensed entities, and an agent holds no licence. Until one of those lands, the wallet is a well engineered way to move a purchase decision into a chat window. It will not change how people pay until a network moves the loss. Everything shipped in 2026, including this wallet, is preparation for that decision, with the risk still sitting where it has always sat: on the merchant’s books, in a dispute queue that has no special code for what just happened. Methodology Data as of November 2026. Launch details come from OpenAI’s announcement as reported by TechCrunch and CNBC on 13 November 2026, including the card networks involved, the token design and the absence of a transaction fee. Market size figures are taken from the published forecasts named in the table, each of which defines agentic commerce differently; they are not directly comparable and are presented side by side rather than averaged. The $823 million figure multiplies EMARKETER’s 2026 estimate by the 4% fee OpenAI charged on Instant Checkout in 2025, which is an illustration of the revenue at stake, not a forecast of what any firm will earn. Consumer trust percentages come from surveys commissioned by Visa and Accenture, with sample sizes of 2,065 and 25,590 respectively. Dispute rules were checked against network programme documentation available in November 2026 and may change without notice. Sources TechCrunch , ChatGPT’s new Wallet lets AI agents shop and check out for you, 13 November 2026 CNBC , OpenAI launches ChatGPT Wallet, 13 November 2026 TechCrunch , United Airlines blocks AI agents from booking flights, 12 November 2026 Retention Edge , AI in ecommerce statistics for 2026, accessed November 2026 Payment Review , a year of agentic checkout, November 2026 About Global Infinity Labs Global Infinity Labs is an AI innovation company specializing in MCP data services. This article was prepared by our expert team using insights and data provided through our MCP platforms. Contact: support@globalinfinitylabs.com
More
Anthropic’s Quarter: $17 Billion In, $9.9 Billion Out to Compute SeaPRwire

Anthropic’s Quarter: $17 Billion In, $9.9 Billion Out to Compute

The amended S-1 puts third quarter compute at 58% of revenue, with $895 billion of purchase commitments against an $81 billion run rate Anthropic’s amended prospectus, filed on 2 November 2026, added one quarter to the numbers investors had already seen. The quarter arrived with $16.96 billion of revenue, up from $11.56 billion in the previous three months and up 1,127% from a year earlier. It also arrived with an operating loss of $567 million, which reversed the $974 million operating profit the company had reported for the second quarter. The swing is small next to the revenue. It is not small next to the story. Anthropic went into its listing as the AI lab with the cleanest business model, the one selling to enterprises rather than chasing consumers. A quarter that flips from profit to loss because of how a company bought its electricity is a different kind of fact about that business. The filing is unusually specific about why. Of the compute delivered in the three months to September 2026, $5.0 billion came through short-term commitments, which the company says cost significantly more than comparable capacity bought under multi-year contracts. Add $2.0 billion of non-cash charges tied to its employee donation matching programme, and the operating line goes negative. The quarter flipped on $5 billion that arrived late Anthropic explains the short-term buying as a timing failure, not a strategy. Capacity under long-term contracts is typically signed 12 to 18 months before it shows up. When customer demand runs ahead of that pipeline, the company either waits or pays the spot price. In the third quarter it paid. The premium it paid is not disclosed in dollars. What is disclosed is the effect on the ratio that matters: revenue generated per dollar of compute fell to 1.72x, after peaking above 2x in the previous quarter. If the short-term capacity had cost the same as contracted capacity, the operating line would have been positive. Two details suggest this is manageable rather than terminal. Adjusted operating income, which strips out share compensation and the donation charge, rose to $2.46 billion from $2.23 billion, though the margin fell to 15% from 19%. And the customer base is broadening: 79% of third-quarter revenue came from outside the five largest customers, up from about 64% in the first quarter. The company also reported more than 11,900 customers spending over $100,000 a year, with a net revenue retention rate near 900%. A business with that retention can absorb a bad quarter of procurement. It cannot absorb a structural rise in the price of the input that dominates its cost base. Compute went from 251% to 48% to 58% The filing puts the whole trajectory in one series. Compute and infrastructure spending was 251% of revenue in the first quarter of 2025, when the company was small and building ahead of demand. It fell to 48% by the second quarter of 2026, as revenue grew faster than committed capacity. It rose again to 58% in the third quarter. That last move is the one investors will read twice. At the exact moment a company normally shows improving unit economics before a listing, Anthropic showed the opposite. The filing does not hide it. It says the company expects compute and infrastructure to remain its largest operating expense for the foreseeable future, in absolute dollars. It adds: “We expect our use of short-term compute to continue to increase in the near term and to in turn compress our operating margin.” Table 1: The quarterly ladder disclosed in the amended filing Measure Q3 2025 Q2 2026 Q3 2026 Revenue ($bn) 1.38 11.56 16.96 Compute and infrastructure ($bn) not disclosed 5.55 9.86 Compute as share of revenue not disclosed 48% 58% Operating result ($m) not disclosed +974 -567 Adjusted operating margin not disclosed 19% 15% Source: Anthropic Form S-1/A, 2 November 2026, as reported by WebProNews, with second quarter compute dollars derived from the disclosed ratio The 58% is the number to hold on to. It means that for every dollar a customer pays Anthropic, 58 cents goes out the door to compute and infrastructure before a single salary, sales commission or research project is funded. The company still earns an adjusted margin, because the compute number includes depreciation and prepaid capacity timing. But the gross spread is thin for a software business, and it is the spread that decides what the company is worth. Eleven years of revenue sits in the commitments note The commitments section is where the filing stops describing a software company. Anthropic had signed purchase and lease agreements worth about $895.4 billion by 30 September 2026, substantially all of it for compute and infrastructure. Of that total, roughly $69.8 billion falls due within 12 months of 30 September 2026, and about $447.5 billion falls due before the end of 2030. Put those against the run rate. The November filing reports annualised revenue of about $81 billion. That means the total commitment equals roughly 11 years of revenue at the current pace. The near-term slice alone, $69.8 billion, is about 86% of a full year of revenue at that run rate. The company is not insolvent. It ended September 2026 with $146.6 billion of liquidity, including $95.6 billion in cash and $36.3 billion in short-term investments, and it has up to $77 billion of note financing capacity it has not drawn. Nine months of operating cash flow came in at $6.6 billion, which is positive and growing. But the commitments have a shape that matters. For every dollar of revenue the business generates this year, it has already promised more than a dollar of future compute spending. If the 58% intensity holds, absorbing the $447.5 billion due by 2030 requires roughly $772 billion of cumulative revenue over the same period. That is nearly ten times the current annual run rate, compounded over four years. That is not a prediction of failure. It is the arithmetic that the IPO price has to justify. The nine month view behind the quarter One quarter can mislead, so the nine month column is worth reading first. Revenue for January to September 2026 reached $33.2 billion, against $2.5 billion in the same period of 2025. The operating loss for those nine months was $1.26 billion, which means the business was close to break even across three quarters and lost money in only one of them. The net loss tells a different story, and almost none of it is operational. Anthropic reported a net loss of $104.04 billion for the nine months, including a $23.29 billion net loss in the third quarter alone, of which $21.9 billion came from marking convertible notes and similar instruments to a higher valuation. That is an accounting consequence of the company becoming more valuable, and it will not repeat once the notes convert. The cash flow statement is the most useful part for anyone outside the company. Operating activities generated $6.6 billion over the nine months, up from a $2.0 billion outflow a year earlier. Investing activities consumed $36.5 billion, mostly short-term investments, lease prepayments and capital expenditure. Financing brought in $110.7 billion, overwhelmingly preferred stock. In plain terms, the business now funds its own operations and still needs investors to fund its buildings. That split is why the listing matters more than the loss. Anthropic’s operating cash flow covers the lights. It does not cover $447.5 billion of commitments due by 2030, and equity is the only instrument large enough to close that gap. Renting at the top of the market Where the compute comes from explains the price. Anthropic has been paying SpaceX $1.25 billion a month for capacity through May 2029, alongside multi-billion dollar deals with Nscale, Akamai and Volta. These are neoclouds rather than hyperscalers: smaller operators that sell flexible capacity at higher rates, sometimes in buildings they lease themselves. The trade is speed for margin. A hyperscaler contract priced 18 months ahead is cheap and inflexible. A neocloud contract is expensive and can be switched on in weeks. When demand doubles in a quarter, the second option is the only one that works. The awkward part is that the largest cloud providers are also becoming competitors. Reporting on the S-1 said Anthropic had struggled to secure enough capacity from AWS, Google and Microsoft, which increasingly view the Claude maker as a rival rather than a customer. The company’s own filing warns that the third-party data centre operators it leases from have a limited operating history. Turbines are the same decision in physical form. In November 2026 Anthropic agreed to buy about $3.4 billion of Caterpillar gas generators, capable of producing up to 1.7 GW, from equipment that the supplier OnePWR had ordered years earlier and could not finance. Paying for someone else’s abandoned order is how you skip the queue for grid connection. Anthropic Direct is the bet that fixes the margin The filing names the mechanism the company intends to use to repair its compute margins: Anthropic Direct, a programme that leases chips and hosts them in dedicated facilities. The logic is straightforward. Instead of buying capacity from a neocloud at spot rates, the company takes the equipment and the building on long leases and operates them itself, which converts an expensive variable cost into a cheaper fixed one. The filing says the programme will let it backfill expensive short-term capacity as its own infrastructure comes online. That is the same bet every large lab is making, and it carries the same risk. A fixed lease is cheaper per GPU hour only if utilisation stays high for the length of the lease. If demand softens in 2028, the company is left paying for capacity it does not need, which is exactly the position the short-term purchases were meant to avoid. The difference between a good and a bad outcome is a utilisation forecast that nobody has published. The product mix pushes in the opposite direction. In November 2026 alone, Anthropic shipped Claude Meetings, which turns a live discussion into a plan and a task list, and a personal finance feature that connects Claude to bank accounts through Plaid. Both are agentic: they run continuously, call tools and keep context. Work of that shape consumes far more compute per user than a chat window, which means the intensity number is more likely to rise than fall. The turbines cost $2,000 per kilowatt That deal is worth pricing per unit, because it shows what urgency costs. For delivery from 2027, $3.4 billion for 1.7 GW works out to about $2,000 per kilowatt of generating capacity, for equipment only, before installation, fuel and grid interconnection. New combined cycle gas plants in the Gulf are being built for roughly $1,150 per kilowatt of installed capacity, based on the $3 billion budget for the 2.6 GW Taweelah C project in Abu Dhabi. The comparison is not like for like, since Anthropic is buying simple cycle generators rather than a combined cycle plant, and simple cycle costs less per kilowatt to build but burns more fuel per megawatt hour. Even allowing for that, the premium is for time. The global backlog for gas turbines runs into the early 2030s, and grid connection queues are just as long. A company that expects to be serving $81 billion of annual demand cannot wait five years for a slot, so it pays for somebody else’s. The fuel bill is the part that stays. Simple cycle generators run at roughly 35% thermal efficiency, against 60% for a combined cycle plant. Every megawatt hour from those turbines costs more in gas than the same hour from a modern plant, and gas prices in some US markets are forecast to more than triple by the end of the decade. The $3.4 billion buys the capacity. It does not fix the operating cost. What has to be financed before 2030 Bring the pieces together and the shape of the company becomes clear. In the twelve months to September 2026, Anthropic turned a research lab into a business with an $81 billion run rate, 11,900 customers above $100,000 a year and a retention rate near 900%. In the same period it signed commitments equal to 11 years of that revenue, borrowed nothing, and funded itself with preferred stock. That combination works while equity markets are willing to price growth. It becomes uncomfortable the moment a quarter of margin compression coincides with a credit rating that limits borrowing. The $567 million loss in the third quarter is a detail. The 58% compute intensity and the $69.8 billion due within a year are the terms of the business. What the listing prices The IPO terms put the company at $70 to $78 a share, or roughly $1.27 trillion to $1.45 trillion on a fully diluted basis, with 825 million shares on offer and a separate $26 billion private placement involving Google, Microsoft, Nvidia, AMD, Broadcom and Micron. The shares are expected to start trading on Nasdaq on 19 November under the symbol ANT. At the November 2026 run rate of $81 billion, a $1.4 trillion valuation is about 17 times sales. That is a growth multiple, not a margin multiple, which tells you what buyers are being asked to believe: that revenue keeps compounding at a rate that lets a company with a 15% adjusted margin and 58% compute intensity grow into its commitments. Credit agencies are less romantic. The Financial Times reported that Anthropic is expected to receive a high junk rating after listing, which would make future borrowing more expensive. A company that has promised $895 billion of compute spending and carries a junk rating has one source of funding left at scale: its own equity. The filing also shows the price of competing on price. Reductions introduced with new model versions and contractual discounts cut about $600 million from third-quarter revenue in 2026, roughly 3.5% of the total. Discounting to hold share against OpenAI and Google is a choice the company can make while it is raising money. It is harder to sustain once the money is raised and the commitments come due. Three blanks left in the filing Three things would change the picture, and none of them is disclosed yet. The premium Anthropic pays for short-term compute, in dollars per GPU hour. Without it, nobody outside the company can model how much of the margin compression is temporary. The mix between owned capacity and rented capacity by 2028. Owned capacity carries depreciation and capital cost; rented capacity carries the neocloud margin. The two produce very different margins on the same revenue. The run rate as of 31 October, which the amendment left blank. Every valuation comparison in the market depends on that number and the filing stops at 30 September. Until those arrive, the defensible reading is narrow. Anthropic has built a business that grew revenue 1,127% in a year, retains customers at a rate software companies rarely see, and just discovered that its largest cost is also its least predictable. The $567 million loss is small. The $895 billion commitment is not. Computation Data as of November 2026. All figures come from Anthropic’s amended Form S-1 filed with the SEC on 2 November 2026, as reported by WebProNews, FourWeekMBA and Caixin, or from the original S-1 filed on 26 October. Second quarter compute spending of $5.55 billion is derived by applying the disclosed 48% ratio to disclosed revenue, because the filing gives the ratio rather than the dollar figure. The $772 billion cumulative revenue estimate divides the $447.5 billion commitment due by 2030 by the disclosed 58% compute intensity and holds that intensity constant, which is an assumption, not a forecast. The $2,000 per kilowatt figure divides the $3.4 billion turbine purchase by 1.7 GW of capacity and excludes installation and fuel. Taweelah C’s $1,150 per kilowatt uses its $3 billion budget and 2.6 GW nameplate. Sources WebProNews , Anthropic brought in nearly $17 billion and slipped back to an operating loss, 2 November 2026 FourWeekMBA , twelve new numbers in Anthropic’s S-1/A, 2 November 2026 TechCrunch , Anthropic is still grappling with the economics of compute, 3 November 2026 TechCrunch , Anthropic buys $3.4B worth of turbines, 9 November 2026 TechCrunch , Anthropic expected to receive a high junk rating, 30 October 2026 About Global Infinity Labs Global Infinity Labs is an AI innovation company specializing in MCP data services. This article was prepared by our expert team using insights and data provided through our MCP platforms. Contact: support@globalinfinitylabs.com
More
E-Commerce Ad Costs Head Into Black Friday With a 14% Price Ramp SeaPRwire

E-Commerce Ad Costs Head Into Black Friday With a 14% Price Ramp

Meta’s third-quarter report, released on 21 October 2026, put a number on the squeeze that e-commerce advertisers have been feeling all autumn: ad impressions delivered across the Family of Apps rose 10 percent year over year, and the average price per ad rose 14 percent in the same period. The auction cleared more inventory and charged more for it at once, with total revenue of 64.60 billion dollars, up 26 percent year over year, and 3.64 billion daily active people. For sellers budgeting campaigns against Black Friday on 27 November, the working question is what each placement actually costs in late 2026 and which parts of the bill respond to planning rather than luck. The clearest read on late-2026 pricing comes from Emplifi’s quarterly benchmarks, published on 12 November with third-quarter data. Across Facebook and Instagram ads in every industry, the worldwide click-through rate reached 1.28 percent in September 2026 at a median cost per thousand impressions of 2.25 dollars. In the United States, click-through rates rose 30 percent year over year to 2.05 percent, with September CPMs at 8.53 dollars, and Facebook feed stayed the cheapest placement for clicks at 0.27 dollars per click on a 1.67 percent click-through rate. Those medians cover all industries, so an online store running purchase campaigns should expect to sit above them. Placement-level numbers show how wide the spread is inside one platform. Emplifi’s third-quarter placement benchmarks put Facebook feed at a 3.94 dollar CPM and Facebook Reels at 4.17, with Instagram Reels at 4.22, Instagram Stories at 5.14 and Instagram feed at 5.18 across global advertising accounts. E-commerce ad accounts pay more per impression on Instagram: medians of 5.54 dollars in feed, 5.50 in Stories and 4.28 in Reels, with click costs between 0.54 and 0.81 dollars. A dataset covering more than 40,000 Meta advertisers from August 2025 to July 2026 puts the median e-commerce CPM at 15.06 dollars, ranging from 11.50 dollars in baby products to 21.80 dollars in health and wellness, with a median cost per purchase of 38.99 dollars and a median click-through rate of 2.39 percent. The category spread on purchases is wide enough to break national budget templates: the same dataset places median cost per purchase near 27 dollars in e-learning, around 30 dollars in baby products and up to 52 dollars in electronics, and sixteen of the seventeen industries in it paid more for reach than a year earlier, with median CPM up about 13 percent year over year. The year-over-year comparison inside the click benchmarks is more forgiving than the impression side. The 2026 edition of the WordStream click dataset, covering April 2025 to June 2026 with 1,377 traffic campaigns and 452 lead campaigns, put median traffic CPC at 0.60 dollars against 0.70 dollars in the 2025 edition, and median lead CPC at 1.80 dollars against 1.92, with cost per lead effectively flat at 27.39 dollars against 27.66. Clicks got cheaper while impressions got dearer because more of the auction shifted toward optimization for outcomes rather than raw delivery, which is exactly the shift a seller can plan around. The picture beyond Meta is where budgets get misallocated. Google search clicks on brand terms ran at a median 1.02 dollars against 2.14 dollars for non-brand search in a panel of 28 DTC brands with 4.36 million dollars of managed spend, which means the same budget buys half as many non-brand visits. TikTok’s median CPM sat at 4.08 dollars with a 0.61 percent click-through rate and a return on ad spend of 1.51 in the same research series, against 2.55 for Meta, so its cheap reach does not carry purchases at the same rate. Lead generation on Meta runs on different economics entirely: the 2026 WordStream benchmarks, covering April 2025 to June 2026, put lead campaigns at 1.80 dollars per click and 27.39 dollars per lead, with dentist campaigns at 61.56 dollars per lead and real estate at 13.74. These numbers only reconcile if each one is judged by the outcome it buys. A 0.27 dollar Facebook feed click and a 2.14 dollar non-brand search click are not substitutes: one interrupts browsing, the other answers an active query. Cost per thousand impressions is the right lens for awareness, cost per click for traffic, and cost per purchase for everything that pays the bills, because a platform with expensive clicks and high intent can still deliver a lower cost per acquisition than a cheap-click, low-intent one. Mixing those three lenses in one budget review is the most common planning error in Q4. Seasonality adds a predictable layer. Cost per thousand impressions on Instagram climbed from 6.21 dollars in January to 8.80 dollars in November in prior-year tracking, roughly 40 percent across the year with the steepest stretch after October, and Black Friday week has historically been the most expensive buying window of the calendar. That lands on top of demand-side pressure: Meta guided fourth-quarter revenue to a range of 73 to 76 billion dollars, implying continued auction competition into January. Emplifi’s survey of 1,638 shoppers in the US and UK found 89 percent rate free shipping as important and 86 percent start buying gifts by Black Friday, so the week both costs the most and converts the best. The audience side of the same benchmarks sets the creative agenda for that week. A typical online store’s Instagram post earned 28 interactions in September against 3 on Facebook, with Reels at 32 median interactions and carousels at 29 against 15 for a single image, roughly double per post. Delivery expectations sit alongside the discount: 73 percent of surveyed shoppers would pay extra for guaranteed delivery dates, and 88 percent expect a reply within 24 hours during peak, which turns saved replies and weekend inbox cover into revenue protection rather than customer-service overhead. Stores that treat those two numbers as campaign inputs, not as afterthoughts, hold conversion while their competitors bid against each other on price alone. Several inputs that inflate the bill are controllable. Meta’s learning phase needs about 50 optimization events per ad set within seven days, which at a 30 dollar cost-per-purchase target means roughly 215 dollars a day per ad set to exit learning inside a week; thinner budgets simply stay in learning longer and pay auction prices without accumulated signal. Creative volume matters because ad sets with refreshed native video and broad audiences clear at lower prices than fatigued static campaigns, and privacy-related signal loss inflates the measured cost per result even when the true outcome has not moved: consent banners and ad blockers cut the conversion data the auction receives, so the reported cost per purchase drifts upward while the actual purchase rate holds. A budget review that separates reported cost from auction cost, then rebuilds events through server-side tracking, routinely recovers several points of measured performance before a single bid change. The constructive read of a 14 percent annual price ramp is that it prices out weak measurement faster than it prices out small budgets. Click-through rates rising 30 percent year over year in the US show that relevance still earns cheaper delivery, and placement spreads of a few dollars between feed, Stories and Reels reward routing traffic campaigns to the cheapest click and purchase campaigns to the strongest intent. A 30 to 100 dollar daily test budget on Meta, a feed split that keeps prospecting on Reels and Stories while retargeting holds the feed, and a quarterly reconciliation of cost per purchase against the 38.99 dollar median turn a rising auction from a headwind into a discipline check that most competitors in any given category have not yet run. Sources Meta, “Meta Reports Third Quarter 2026 Results” (21 October 2026): ad impressions up 10 percent year over year, average price per ad up 14 percent, revenue 64.60 billion dollars up 26 percent, 3.64 billion daily active people, and fourth-quarter revenue guidance of 73 to 76 billion dollars. Emplifi, “How to prepare your social media for Black Friday (with Emplifi peak season data)” (12 November 2026): worldwide CTR of 1.28 percent in September 2026 at a 2.25 dollar CPM, US CTR up 30 percent year over year to 2.05 percent with September CPMs at 8.53 dollars, Facebook feed at 0.27 dollars per click, and e-shop and consumer survey benchmarks. TrackBee, “Ad Cost Benchmarks 2026” and “How Much Do Facebook Ads Cost in 2026”: e-commerce benchmark of more than 40,000 Meta advertisers from August 2025 to July 2026 (median CPM 15.06 dollars, CPA 38.99 dollars, industry spread 11.50 to 21.80 dollars), Meta learning-phase requirements, and Emplifi Q3 2026 e-commerce placement medians. Top Growth Marketing, “DTC and eCommerce Benchmarks”: first-party medians across 19 DTC brands and 2.88 million dollars of Meta spend (CPM 13.40 dollars, ROAS 2.55), 28 brands and 4.36 million dollars of Google spend (non-brand search CPC 2.14 dollars, brand 1.02 dollars), and the TikTok panel (median CPM 4.08 dollars, ROAS 1.51). PPCone, “How much do Facebook and Instagram ads cost”: Emplifi Q3 2026 placement benchmarks (Facebook Feed CPM 3.94 dollars, Facebook Reels 4.17, Instagram Reels 4.22, Instagram Stories 5.14, Instagram Feed 5.18) and WordStream 2025 versus 2026 edition comparisons. Mark Studios, “How Much Do Instagram Ads Cost in 2026”: LocaliQ 2026 benchmarks published with WordStream (lead campaigns at 1.80 dollars per click, 27.39 dollars per lead, industry spread from 61.56 dollars for dentists to 13.74 dollars for real estate) and Q3 2026 account-level placement data. Disclaimer: This article is provided for general information only and does not constitute legal advice. About Global Infinity Labs: Global Infinity Labs is an independent research desk covering cross-border commerce, logistics and regulation. The desk publishes field notes for operators, compliance teams and investors working across Asian and Gulf markets. Email: support@globalinfinitylabs.com Submission portal: https://newsroom.seaprwire.com
More
US Sales Tax Clocks Diverge After the 2026 Nexus Changes SeaPRwire

US Sales Tax Clocks Diverge After the 2026 Nexus Changes

A seller crossing Illinois’ 100,000-dollar threshold in late November faces a different compliance clock from one crossing North Carolina’s threshold the same week. Since 1 January 2026, Illinois has evaluated its threshold at the end of every calendar quarter and automatically reclassified registered sellers who slip under it into voluntary use-tax filing. North Carolina, since 2 July 2026, gives threshold-only sellers at least 60 days before the collection duty begins. Both rules took effect this year, and together they show where United States economic nexus enforcement is heading: fewer transaction counts, tighter dollar-line tracking, and registration deadlines that no longer share a single date across the 45 states that levy sales tax. Illinois rebuilt its rule through Public Act 104-0006, which eliminated the 200-transaction test for remote retailers and marketplace facilitators on 1 January 2026 and left 100,000 dollars in cumulative gross receipts over a preceding 12-month lookback as the sole trigger. The lookback is evaluated at the end of every calendar quarter, on 31 March, 30 June, 30 September and 31 December. A registered seller whose trailing receipts fall under the threshold must stop collecting destination-based retailers’ occupation tax, and the state did not close those accounts: it converted them to voluntary use-tax status collecting the flat 6.25 percent state rate, so a seller that wants out of filing entirely has to cancel the registration deliberately. Sales for resale backed by a valid certificate and sales made through a marketplace facilitator that certifies collection are excluded from the threshold calculation. Illinois also ran a one-time cleanup for the transition period. The 2026 Remote Retailer Tax Amnesty Program, announced on 5 June 2026 and running from 1 August to 31 October 2026, let remote retailers settle eligible liabilities for sales from 1 January 2021 to 30 June 2026 at simplified rates of 9 percent for general merchandise and 1.75 percent for qualifying food and medicine, with penalties and interest waived. Registration processing takes one to two business days, and from the November 2026 reporting period onward the standard rules apply, which makes the amnesty the boundary between a forgiving registration path and ordinary audit exposure. The transaction-count retreat is broader than one state. Kentucky removed its 200-transaction test on 1 August 2026, and as of that date the states identified as having eliminated the count test include Illinois, Indiana, Iowa, Kentucky, Maine, North Dakota, North Carolina, South Dakota, Utah, Washington, Wisconsin and Wyoming. Ohio still applies a 100,000-dollar sales threshold or 200 separate transactions in the current or prior calendar year, and New York keeps a dual test of 500,000 dollars and 100 sales where both must be met, so a national compliance calendar built on one trigger formula does not survive contact with the state map. The 2026 retreat from transaction counts has a history worth keeping straight. The United States Supreme Court upheld South Dakota’s remote-seller law in South Dakota v. Wayfair on 21 June 2018, ending the physical-presence rule, and New York’s threshold took effect the same day; New York’s marketplace rule followed on 1 June 2019, and Missouri became the last of the 45 sales-tax states to adopt economic nexus and marketplace rules on 1 January 2023. Every state has now held the framework for at least three full years, which is why the current wave of changes is about mechanics, such as when the clock starts and what counts toward the threshold, rather than about whether remote sellers owe the tax at all. North Carolina moved the clock instead of the threshold. Under Session Law 2026-31, effective 2 July 2026, a remote seller whose only connection to the state is exceeding the 100,000-dollar gross-sales threshold becomes engaged in business on the first day of the first calendar month that begins at least 60 days after crossing, and must register and start collecting on or before that date. The department’s notice walks the math through two examples: a seller crossing the threshold on 5 July 2026 registers and collects by 1 October, and a marketplace facilitator crossing on 27 November, Black Friday this year, registers by 1 February 2027. The window exists only for threshold-only sellers; inventory in the state, employees or representatives start the obligation immediately, and sellers who crossed before 2 July 2026 stay under the prior law. The threshold math itself is broad: taxable, exempt, resale, nontaxable and marketplace-facilitated sales all count. Alaska ran its own 2026 change through local administration rather than state statute. Under Juneau Ordinance 2026-22, passed on 8 June 2026, registration, reporting and administration of qualifying marketplace facilitator sales tax moved from the City and Borough of Juneau to the Alaska Remote Seller Sales Tax Commission on 1 October 2026. The commission’s threshold is 100,000 dollars in gross sales delivered into Alaska in the current or previous calendar year, measured statewide rather than by city, with the 200-transaction test removed since January 2025 and marketplace sales counted in the facilitator’s threshold. Sellers meeting it must register within 30 days; sellers whose Alaska sales run entirely through a facilitator generally file a marketplace seller affidavit instead, and facilitators registered directly with Juneau close those accounts and move to the commission’s portal. The rest of the map varies on both dials. California and Texas set their thresholds at 500,000 dollars, Alabama at 250,000 dollars measured on the previous calendar year’s sales, and most other states hold at 100,000 dollars. The registration clock ranges from immediate to four months: Kansas requires collection beginning with the next transaction after the threshold is crossed, New Jersey allows up to 30 calendar days, Virginia requires registration within 30 days of nexus, South Dakota uses the first day of the month at least 30 days after crossing, Tennessee starts the duty in the third calendar month, and Texas generally starts it no later than the fourth month after a seller exceeds its 500,000-dollar safe harbor. Marketplace treatment splits the map again: California counts marketplace-facilitated sales toward an individual seller’s own threshold, while states including Alabama, Arizona, Colorado, Florida, Illinois, Indiana, Maine, Massachusetts, New Mexico, Oklahoma, Tennessee, Utah, Virginia and Wyoming exclude certain marketplace sales from it. For cross-border sellers the practical sequence is the same regardless of which clock applies. Track gross receipts sourced to each state monthly rather than annually, record the exact date a threshold is crossed, check for physical presence through inventory storage, trade-show activity or employees before assuming the grace window exists, and register ahead of the computed start date instead of the due date. Registration itself runs through state portals, with Illinois filing through MyTax Illinois, North Carolina accepting its online registration system or the Streamlined Sales Tax Registration System, and the Alaska commission registering sellers through its filing portal within the 30-day window. Physical presence deserves its own line in the review because fulfillment networks move inventory automatically: stock placed in a state’s fulfillment center can create a collection obligation at any sales level, and the 60-day and 30-day grace windows exist only for sellers whose sole connection is the dollar threshold. Two clarifications keep the calendar honest. First, marketplace facilitator collection does not erase a seller’s own tracking duty, because several states count marketplace sales toward the seller’s threshold even while the platform remits the tax. Second, federal reporting is a separate track entirely: third-party settlement organizations report payments on Form 1099-K when gross payments exceed 20,000 dollars and 200 transactions in a year, that is a reporting threshold rather than a tax determination, and the form shows gross amounts without netting out fees, refunds, shipping or discounts, so it reconciles against the seller’s own books rather than against state registration duties. The 2026 direction favors sellers on paper: simpler triggers, explicit grace periods and a forgiving Illinois amnesty window. What the simplification costs is precision, because dollar lines rather than order counts now decide when the duty started, and a fourth-quarter surge that crosses a threshold in mid-December starts a clock whose first day lands in a different quarter entirely. Sources North Carolina Department of Revenue, “Important Notice: Extended Compliance Period for Certain Remote Sellers” (6 August 2026): Session Law 2026-31 Section 3.1 effective 2 July 2026, the 60-day rule for threshold-only remote sellers, official examples crossing the threshold on 5 July 2026 and 27 November 2026, and the physical-presence exclusion. Illinois Department of Revenue, “FY 2026-28, 2026 Illinois Remote Retailer Tax Amnesty Program”: amnesty running 1 August to 31 October 2026, eligible liabilities for sales from 1 January 2021 to 30 June 2026, simplified rates of 9 percent and 1.75 percent, registration and MyTax Illinois requirements, and the threshold mechanics under FY 2026-12. State of Illinois, “Illinois Department of Revenue Announces 2026 Remote Retailer Tax Amnesty Program” (news release): official program announcement dated 5 June 2026 with penalty and interest relief, streamlined reporting and simplified tax rates. Sterlinx Global, “USA Update: Alaska Moves Juneau Marketplace Facilitator Tax to the Statewide ARSSTC System”: Juneau Ordinance 2026-22 passed on 8 June 2026, the 1 October 2026 transfer of marketplace facilitator administration to the Alaska Remote Seller Sales Tax Commission, the 100,000-dollar statewide Alaska threshold, and the 30-day registration requirement. Alaska Remote Seller Sales Tax Commission, official information portal: intergovernmental agreement of 2019, business and seller registration information, and remote sales tax administration for participating local jurisdictions. OurTaxPartner, “Goodbye 200-Transaction Rule: How Economic Nexus Simplified in Illinois, Utah, and Alaska for 2026-2027”: Illinois Public Act 104-0006, the quarterly 12-month lookback evaluated at calendar quarter ends, automatic voluntary use-tax reclassification, resale and marketplace exclusions from the threshold, and the amnesty boundary against voluntary disclosure rules. Source Advisors, “Economic Nexus in 2026: Why a 100,000 Threshold Does Not Mean the Same Thing Everywhere”: Kentucky’s transaction-test repeal effective 1 August 2026, Ohio’s retained alternative test, state-by-state registration timing from Kansas through Texas, and threshold calculation differences including North Carolina’s broad inclusion rule. Disclaimer: This article is provided for general information only and does not constitute legal advice. About Global Infinity Labs: Global Infinity Labs is an independent research desk covering cross-border commerce, logistics and regulation. The desk publishes field notes for operators, compliance teams and investors working across Asian and Gulf markets. Email: support@globalinfinitylabs.com Submission portal: https://newsroom.seaprwire.com
More
Battery Freight Into Saudi Arabia: The Document Chain to Clearance SeaPRwire

Battery Freight Into Saudi Arabia: The Document Chain to Clearance

Saudi Arabia’s updated import licensing guide took effect on 1 November 2026, issued under Ministerial Decision No. 70 and replacing the edition that had run since December 2019. The guide sets out responsible authorities, category-specific conditions, supporting documents, application procedures, processing periods and appeal mechanisms for restricted goods, and restates the border rule in plain terms: any shipment arriving without a pre-approved, valid import permit is rejected at the importer’s sole expense. For freight moving between the United Arab Emirates and Saudi Arabia with lithium batteries on board, that change is the visible tip of a document chain that now decides whether a truck crosses on schedule, sits in Dammam’s inspection zone, or turns back at the border. The conformity layer runs through the SABER platform of the Saudi Standards, Metrology and Quality Organization, an institution whose published counters list 33,817 approved standards and more than 82,000 certificates granted to date. A battery product needs two certificates bound together: the Product Certificate of Conformity, a model-level registration valid for one year, and the Shipment Certificate of Conformity, issued per consignment against the product certificate and required for customs clearance of every batch. Only a registered Saudi importer can file in SABER; the overseas factory cannot register directly, which makes the importer’s account status a pre-shipment check rather than an afterthought. On the standards side, portable cells and battery packs certify against SASO/IEC 62133-2:2017 with a three-year certificate under annual surveillance, stationary and commercial energy storage certifies against SASO/IEC 62619:2022, and the Saudi battery standard SASO 2870 adds tropical-climate testing at 50 degrees Celsius and 95 percent relative humidity plus Arabic-language marking on top of the IEC baseline. Certification service timelines reported by conformity bodies put a product certificate converted from an existing IEC CB report with Saudi national differences at two to four weeks, a full test-and-certify run without a CB report at six to eight weeks, and a shipment certificate at one to three days once the product certificate stands, so the product certificate is the long pole and belongs at the top of the pre-loading checklist. The transport layer is international and does not pause at the border. Lithium cells and batteries must carry a UN 38.3 test summary covering the eight-test sequence from altitude simulation through overcharge, a safety data sheet in both English and Arabic, and the lithium battery mark with the Class 9 label on every package. Standalone lithium-ion batteries ship as UN 3480 and batteries packed with or contained in equipment as UN 3481, both Class 9 with no limited-quantity exemption, so above the small-battery thresholds the cargo is fully regulated in every mode. Lithium-ion accumulators sit under HS code 8507.60, and clearance turns on that description matching across the SABER certificate, the commercial invoice, the packing list and the bill of lading; a certificate that names one thing while the boxes carry another is the classic hold trigger. Watt-hour ratings deserve the same treatment as the HS code, because the rating printed on the cell decides whether the consignment qualifies for the small-battery exemption below or the full dangerous-goods regime above, and a spec sheet from an earlier production run does not travel well through an inspection zone. Road transport has its own granularity. Small batteries ride exempt from full dangerous-goods documentation under special provision 188, aligned in substance across the ADR road code and the IMDG sea code: lithium-ion cells up to 20 watt-hours and batteries up to 100 watt-hours, lithium-metal cells up to 1 gram and batteries up to 2 grams, with the battery mark and no shipper’s declaration when every condition holds. Above those limits the consignment is fully regulated Class 9 cargo in all three modes, and the ICAO Technical Instructions 2025-2026 edition, valid until 31 December 2026, governs any air leg into the region. The Gulf has no single road code equivalent to ADR, so each road leg runs under national dangerous-goods requirements, including civil defense rules on the Saudi side, which is why the truck file needs the same UN 38.3 and marking documentation the sea leg used. The UAE side of the lane adds a second conformity file. Regulated battery products fall under the Emirates Conformity Assessment Scheme, with certificates issued by bodies approved by the Ministry of Industry and Advanced Technology, and ECAS certificates are model- and capacity-specific: a manufacturer-wide certificate covering a product family is not valid for clearance. Storage systems certify against IEC 62619 as the primary standard, with UL 1973 accepted as an equivalent for stationary batteries, performance testing follows IEC 62620, and the UN 38.3 report carries over from the transport file. Grid-connected storage in Dubai additionally requires approval from the Dubai Electricity and Water Authority, and stationary installations face civil defense requirements across the emirates. Packaging and declaration discipline closes the chain on the physical side. Cases need compression strength and short-circuit protection, positive terminals sealed with insulating tape, loose cells packed in individual insulation bags away from metal parts, and warning marks that identify the cargo to every handler along the route. Saudi customs declaration runs through the FASAH window with the HS code, declared value and specifications matching the certificate file, and underspecified descriptions such as generic electronics labels invite manual inspection that adds days a battery shipment rarely has to spare. Enforcement at the Saudi end has sharpened into a pattern worth quoting directly. Trade reporting from Dammam in early November described a 40-foot container of 480 cordless drills held for 11 days in the customs inspection zone because the SABER product certificate described the drill bodies while the lithium-ion packs rode the packing list as separate line items, with demurrage running around 50 dollars per twenty-foot unit per day plus storage, and fines for incomplete dangerous-goods paperwork reported at up to 5,000 Saudi riyals. The recurring trigger in these cases is not a missing test report but a description mismatch across documents, and customs officers are trained to compare the trade description, HS code, SABER registration and dangerous-goods status before releasing a container for inland movement. The sequence that clears is assemble-in-order. Fix the HS code and watt-hour class before quoting freight, because the classification drives both the conformity path and the dangerous-goods status. Collect the UN 38.3 test summary, IEC test reports and quality-management certification from the factory before price negotiations conclude, since a factory without them cannot support the conformity file at any price. File the product certificate before loading and the shipment certificate per consignment against the export file, keep the safety data sheet in both languages, and freeze master data after the shipping-instruction cut-off so certificate, invoice and bill of lading cannot drift apart. On the road leg, dispatch only with the dangerous-goods declaration, marks and labels matched to the mode thresholds above. None of this reads as new regulatory risk, and that is the constructive point for shippers. The 1 November licensing guide shortens processing periods, names the responsible authority per category and sets appeal routes, which cuts ambiguity for compliant filings rather than adding a hurdle. Batteries move into and across the Gulf at scale every day; what the current enforcement wave rewards is document consistency, assembled in the order the chain expects, with the importer’s SABER account confirmed before a container is loaded rather than after it is held. Sources Hong Kong Trade Development Council, “Saudi Arabia: New Import Licensing Guide Released”: the updated guide effective 1 November 2026 under Ministerial Decision No. 70, replacing the 2019 edition, covering responsible authorities, category conditions, processing periods and appeal mechanisms, and the border rejection rule for shipments without pre-approved permits. Hainan Provincial Administration for Market Regulation, “Middle East Energy Storage Market and Compliance Points” (24 August 2026): the Saudi SASO/SABER framework with PCoC and SCoC certificates, SASO/IEC 62133-2:2017 for portable batteries, SASO/IEC 62619:2022 for storage systems, SASO 2870 tropical-climate testing and Arabic marking, and the UAE ECAS regime with model-specific certificates, IEC 62619, UL 1973 equivalence and DEWA approval for Dubai storage. DGAssistant, “Lithium batteries: documents and marks by air, sea and road”: UN 3090/3091/3480/3481 classifications as Class 9 with no limited quantities, special provision 188 thresholds for small batteries, ADR 2025 and IMDG Amendment 42-24 references, and the ICAO Technical Instructions 2025-2026 edition valid until 31 December 2026. Keerki, “Importing EV and Lithium Batteries from China to the Gulf: The Complete 2026 Guide”: UN 38.3 test summary requirements, dangerous-goods declaration and safety data sheet paperwork, container packing practices, and the SABER and ECAS conformity routes for Saudi Arabia and the UAE. Industry trade reporting from Dammam, November 2026: a 40-foot container of 480 cordless drills held for 11 days over a SABER certificate and packing-list description mismatch, demurrage of about 50 dollars per twenty-foot unit per day, fines for incomplete dangerous-goods paperwork reported up to 5,000 Saudi riyals, and HS 8507.60 document consistency checks at Saudi customs. Disclaimer: This article is provided for general information only and does not constitute legal advice. About Global Infinity Labs: Global Infinity Labs is an independent research desk covering cross-border commerce, logistics and regulation. The desk publishes field notes for operators, compliance teams and investors working across Asian and Gulf markets. Email: support@globalinfinitylabs.com Submission portal: https://newsroom.seaprwire.com
More
Manila’s SMX Venues Enter a Back-to-Back Trade Show Stretch SeaPRwire

Manila’s SMX Venues Enter a Back-to-Back Trade Show Stretch

From 25 to 28 November, the SMX Convention Center Manila runs two trade shows on overlapping dates in the same Mall of Asia complex: the 51st edition of 3E XPO, the Philippines’ power and energy exhibition organized by the Institute of Integrated Electrical Engineers, and the Hotel Suppliers Show, the country’s hospitality trade fair managed by Global-Link MP Events International. Both load into the four-hall venue in Pasay within five days of the Philippine SME Business Expo at SMX Aura on 20 November and the DFA International Bazaar at the same property on 29 November. For exhibitors freighting stands into Manila for the year-end stretch, the calendar gives back-to-back intake rather than the clean one-show-in, one-show-out rhythm that single-show planning assumes. The verified sequence reads as follows. Paskomiket, the comics and pop-culture market, filled Hall 1 on 14 and 15 November from 11 in the morning to 8 at night. The 3E XPO then takes the main halls from 25 to 28 November, 9 in the morning to 6 in the evening, with the prior edition covering 20,000 square meters, roughly 300 exhibitors and 25,000 visitors, and a product scope spanning power and energy equipment, smart-grid applications, photovoltaic and storage systems, and microgrid solutions. The Hotel Suppliers Show overlaps from 26 to 28 November across 12,000 square meters, with the last edition counting 255 exhibiting companies and 11,838 attendees, covering hotel furniture, kitchen equipment, guestroom amenities, interior finishes and operating systems for the hospitality trade. SMX Aura, the 26th-floor venue inside SM Aura in Taguig, adds the Philippine SME Business Expo on 20 November from 10 to 7 and the International Bazaar organized for the Department of Foreign Affairs on 29 November from 7 to 7. Across the road, the Philippine International Convention Center carries its own December bookings. PICC, the state-run convention complex with 38 meeting rooms and 70,000 square meters of floor area, has hosted 9,324 events and 460 international summits across five decades of operation, and it remains the venue of record for large official ceremonies in the Bay Area. It hosts the mass oath-taking of new chemists and chemical technicians on 12 December in the Plenary Hall, split into a morning session for chemists and an afternoon session for technicians, with registration deadlines on 4 December. One week later, on 22 December at 9 in the morning, the Professional Regulation Commission’s face-to-face oath-taking for new certified public accountants convenes at PICC, announced on 13 November for the 3,988 passers of the October licensure examination. Both are single-day ceremonies with their own setup and rehearsal days around them. The venue network around the complex is also wider than the two flagship names suggest. SMX Convention Center has operated as the largest privately-owned convention brand in the country since 1992, with Manila joined by SMX Aura in Taguig, SMX Bacolod, SMX Clark and SMX Davao, plus Megatrade Hall and Sky Hall Seaside Cebu under the same management. For a traveling exhibitor the practical effect is that overflow space exists within one contract ecosystem: a booth program that exceeds its main hall allocation can move parallel sessions or storage into an adjacent property under the same operator rather than negotiating with a separate landlord in a different district. December then shifts the complex toward consumer traffic. The AWCP Charity Bazaar runs on 1 December at SMX Aura from 10 to 5, and the World Bazaar Festival, the largest ticketed charity bazaar in the calendar, occupies the World Trade Center Metro Manila in the same Pasay corridor from 11 to 20 December, 10 to 9 daily, organized by Worldbex Services International for the ABS-CBN Foundation. The Noel Bazaar follows in Alabang from 17 to 20 December, and Disney On Ice takes the SM Mall of Asia Arena from 18 December to 3 January. Each of these brings van-scale consumer deliveries and evening teardown loads that trade-show freight never competes with in quieter months. What the overlap means in practice is a compressed loading sequence. Two exhibitions running 26 to 28 November share the same dock windows and freight corridors in the Bay Area on 25 November, when 3E XPO build-up peaks and Hotel Suppliers Show crates arrive behind it. Listed booth rates on the fair listings give a sense of the stakes: standard shell scheme at 3E XPO is quoted around 600 dollars per square meter and space-only at 500, while the Hotel Suppliers Show lists a 9-square-meter package at 5,500 dollars and space-only at 550 per square meter. At those rates, a delayed crate is not an inconvenience; it is a build delay on the most expensive square meters of the year. Freight timing needs to key off venue intake, not off the airport leg. The drive from Ninoy Aquino International Airport to the Mall of Asia complex runs 15 to 30 minutes by taxi or ride-hailing, so the air or sea leg is rarely the binding constraint; dock slots, crate staging space and customs release timing are. Sea freight through the Manila port system or air freight through NAIA both need clearance completed against the build-up date rather than the opening date, because venue intake windows for a 25 November build cannot absorb a container released on the 24th. Space-only exhibitors carry the heavier side of that risk: a bare-plot build waits on its own contractor crew, flooring and rigging deliveries before a shell-scheme booth would, so the crate sequence inside the truck matters as much as the arrival date. Exhibitors booking into the double window should confirm each show’s move-in schedule directly in the exhibitor manual, reserve dock slots as soon as booth contracts are signed, and keep the customs broker working to the same dates. The buyer mix differs across the stretch in ways that change staffing too. Trade-show visitors arrive in badge-wearing waves from the morning opening, while charity and consumer bazaars bring families in continuous flow all day, with peak pressure on parking and ride-hailing stands rather than on freight docks. Event teams moving between the two formats should plan separate crew rosters: dock work concentrates in build-up mornings for the trade shows, and crowd and inbox work concentrates on bazaar weekends in December. The entertainment calendar compounds the pressure on rooms and rides: the Mall of Asia Arena hosts concerts on the first three weekends of December alongside the ice-show run, and hotel occupancy across the Bay Area, Makati and Bonifacio Global City tightens from the second week of the month. Crews arriving for the 26 to 28 November window should book accommodation immediately on booth confirmation, because the same corridors fill with holiday travelers ten days later. Relief is already under construction, which is the constructive note in a crowded calendar. SM Hotels and Conventions has set SMX Seaside Cebu to open in November 2026 with 5.3 billion pesos of investment as the largest upcoming convention venue in the Visayas, three new SMX Trade Halls in Cabanatuan, Santo Tomas and Santa Rosa are slated for fourth-quarter 2026 openings in Luzon, and SMXCITE, a new landmark exhibition center inside the Mall of Asia complex itself, is slated for 2027. Until that capacity lands, the late-November and December calendar in Pasay will keep rewarding exhibitors who treat venue intake, customs release and dock booking as one schedule, planned the day the booth contract is signed rather than the week the crates ship. Sources SMX Convention Center, official venue calendar and press releases: IIEE 3E Xpo 2026 at SMX Manila, NAM Gabi ng Parangal on 20 November, Philippine SME Business Expo and International Bazaar of DFA at SMX Aura on 20 and 29 November, the SMX brand profile, SMX Seaside Cebu opening in November 2026 with 5.3 billion pesos of investment, the SMXCITE project slated for 2027, and three SMX Trade Halls in Cabanatuan, Santo Tomas and Santa Rosa slated for fourth-quarter 2026. What’s On in Manila, “What’s On in Manila: December 2026” (published 4 November 2026): Paskomiket at SMX Hall 1 on 14-15 November, AWCP Charity Bazaar at SMX Aura on 1 December, World Bazaar Festival at the World Trade Center Metro Manila on 11-20 December organized by Worldbex Services International for the ABS-CBN Foundation, Noel Bazaar on 17-20 December, and Disney On Ice at the SM Mall of Asia Arena from 18 December to 3 January. JuFair exhibition listings for 3E XPO 2026 (25-28 November, SMX Convention Center Manila, organized by IIEE, prior edition 20,000 square meters with about 300 exhibitors and 25,000 visitors, listed booth rates) and the Hotel Suppliers Show (26-28 November, organized by Global-Link MP Events International, prior edition 12,000 square meters, 255 exhibitors, 11,838 visitors, listed booth rates, NAIA transfer guidance). Integrated Chemists of the Philippines, “2026 Oathtaking of New Chemists and Chemical Technicians”: 12 December 2026 at the PICC Plenary Hall, morning and afternoon sessions, registration and payment deadline of 4 December 2026. CPA Review PH, “October 2026 CPALE Oath-Taking: December 22 at PICC”: Professional Regulation Commission announcement of 13 November 2026 for the face-to-face mass oath-taking on 22 December 2026 at 9 in the morning, with 3,988 passers out of 10,746 examinees. Philippine International Convention Center, official website: 38 meeting rooms, 70,000 square meters of floor area, and 50 years of operations with 9,324 events and 460 international summits hosted. Disclaimer: This article is provided for general information only and does not constitute legal advice. About Global Infinity Labs: Global Infinity Labs is an independent research desk covering cross-border commerce, logistics and regulation. The desk publishes field notes for operators, compliance teams and investors working across Asian and Gulf markets. Email: support@globalinfinitylabs.com Submission portal: https://newsroom.seaprwire.com
More
Indonesia’s AI gap: 54 percent use it, 14 percent work with it SeaPRwire

Indonesia’s AI gap: 54 percent use it, 14 percent work with it

The numbers say everyone touches AI tools in 2026. The work says few have wired them into the business. Three paths for Indonesian companies, three price tags. An owner of a food business in Jakarta runs the whole company on WhatsApp. Orders arrive there. Suppliers send invoices there. Staff photograph receipts and send them to the same group chat. When someone offers to sell her “AI”, the pitch is usually a chatbot that answers customers faster. That pitch misses the job. The job is the work after the chat: matching invoices to orders, chasing payments, filing records. This gap between talking to AI and working with AI shows up in the national numbers. Indonesia has one of the highest rates of casual AI use in Southeast Asia. It has one of the lowest rates of AI inside business workflows. The distance between those two numbers is where most companies are stuck, and the three ways out of that position cost very different amounts of money and time. The two numbers measure different people The e-Conomy SEA 2025 report, published by Google, Temasek and Bain & Company, puts standalone generative AI use among Indonesia’s working population at 54 percent. The six-country regional average is 58 percent. Thailand sits at 63 percent, Vietnam at 65 percent, Malaysia at 68 percent and Singapore at 70 percent. Indonesia’s internet penetration is 81 percent, so the base is there. The consumer-side numbers run even hotter. The same report says 79 percent of Indonesian users have taught themselves to use AI tools. Veronica Utami, Google’s country director in Indonesia, said at the report launch that 80 percent of users in the country interact with AI daily. She gave one detail that shows the scale: Indonesians generate 18 million images per day on the Nano Banana image feature alone. Revenue from AI-enabled apps in Indonesia grew 127 percent year on year, the fastest in the region. Consumer interest in AI runs about three times the global average, by her account. The enterprise-side numbers tell the other half of the story. Microsoft’s own adoption tracking puts AI use inside Indonesian organizations at 14.1 percent in one measurement period and 14.8 percent in the next. Compare that with the 54 percent individual figure and the ratio is about 3.65 to 1. People play with AI at home. Companies have not rebuilt their work around it. The pattern holds across tools. Claude’s usage tracking shows 23 percent penetration in Indonesian computer and math occupations and 25 percent in education. Lazada reports 31 percent of its Indonesian sellers using its AI sales assistant. These are real numbers, and they are all in the 20 to 30 percent band, far above the 14 percent organization figure and far below the 54 percent consumer figure. The regional spread shows Indonesia is not an outlier in the casual-use number. It is an outlier in the size of the gap. Market Working-population GenAI use Internet penetration Singapore 70 percent Above 90 percent Malaysia 68 percent Above 90 percent Vietnam 65 percent 87 percent Thailand 63 percent Above 90 percent Indonesia 54 percent 81 percent Philippines 53 percent 70 percent Source: Google, Temasek and Bain e-Conomy SEA 2025 , as reported October 2026. Where AI actually touched Indonesian commerce in 2025 was not the office. It was the selling floor. Video commerce, the format where sellers stream and sell inside chat-adjacent apps, grew its active seller count 75 percent year on year to 800,000 sellers. Transaction volume rose 90 percent to 2.6 billion transactions. The overall digital economy is closing on 100 billion dollars in gross merchandise value, with e-commerce at 71 billion dollars. Those transactions increasingly ride on AI-adjacent rails: recommendation feeds, AI-dubbed product videos, automated customer replies. The consumer adoption number and the commerce number describe the same people. The surveys behind these numbers measure different populations, so the exact gap between 54 and 14 percent is an estimate rather than a measurement. Estimates vary on where casual use ends and operational use begins. The direction does not. Casual use runs far ahead of operational use, and the gap is wider in Indonesia than in most of the region because the consumer base grew so fast. Why the obvious answer stalls A company that wants to close this gap usually starts the same way. Someone buys subscriptions to a global AI tool, hands them out, and waits for productivity. In Indonesia this stalls for three reasons that stack on top of each other. The first is language. Global models speak English first. Bahasa Indonesia works, but the languages of daily commerce do not. A shop owner in Surabaya who answers customers in Javanese or Sundanese cannot hand those conversations to a model that fumbles both. The e-Conomy data shows consumer AI activity concentrated in exactly these local-language interactions, which is why adoption exploded among individuals and crawled inside firms. The second is money in the wrong currency. Global AI subscriptions bill in US dollars, typically 20 to 30 dollars per user per month. For a 50-person company that is 12,000 to 18,000 dollars a year before any integration work. That is not a large number by global standards. Against Indonesian SME margins it competes directly with payroll. The third reason is legal, and it arrived on a schedule. Indonesia’s Personal Data Protection Law, UU PDP 27/2022, took full effect in October 2024. Its implementing regulation GR 33/2026 requires controllers to keep a register of processing activities, run impact assessments before higher-risk processing, and honor the right to object to automated decisions. Cross-border transfers of personal data need a legal basis, and remote access from outside Indonesia counts. The Otoritas Jasa Keuangan has issued AI governance guidance for banks, and the regulator treats customer data inside AI prompts as personal data like any other. Enforcement is no longer theoretical. In November 2026, the coffee chain Fore Kopi reported a leak of 5.7 million customer accounts to Komdigi, the communications ministry. The company filed the report itself and apologized publicly. Whatever the investigation finds, the case lands in a system that now has a working complaint channel and a 72-hour breach notification rule for controllers. None of this makes global AI tools illegal. It makes them a compliance project. Sending customer chat histories through a model hosted in another country now needs a documented lawful basis, a transfer mechanism and an impact assessment. For a bank, the review alone can take six months. For a two-person trading company, it takes either a lawyer they cannot afford or a shortcut they should not take. Path one: global subscriptions, kept in their lane The cheapest path is also the most common. A company buys seats on a global tool and uses it where the compliance surface is thin: internal English-language work, drafting, summarizing, coding, analyzing its own non-personal data. The cost is known and small. A team seat runs 25 to 30 dollars per user per month at current list prices. Deployment takes an afternoon. The tools are the best in the world at general reasoning and English output. For a company whose staff already read and write English at work, this path delivers value in week one. The lane restrictions are the whole point. Customer data stays out. Personal data stays out. Anything the PDP Law would call processing of personal data stays out, because the models sit on servers in other countries and no company wants to build a cross-border transfer file for a chatbot subscription. One Jakarta bank described in a 2026 industry guide kept GPT-4o for internal English tooling while moving its customer-facing Bahasa chatbot to local infrastructure. That split is the pattern the market found: global tools for the back office, local systems for the front line. The weakness of this path is that it barely touches the 14 percent number. Internal drafting makes individuals faster. It does not automate invoices, reconcile orders or answer Javanese customers. Companies that stop here stay stuck between the two adoption curves, paying a new subscription and reporting no change in output. Path two: the sovereign hosted route The second path is new in 2026. Sahabat-AI is the open-source model family built by GoTo and Indosat Ooredoo Hutchison, now at 70 billion parameters, running on Lintasarta’s GPU Merdeka infrastructure inside Indonesia. The project had a public origin worth recalling: NVIDIA’s chief executive introduced it at Indonesia AI Day in 2024, the model family reached 70 billion parameters by mid-2025, and the consumer platform arrived in early 2026. The public chat service is free, on the web and inside the GoPay app. The government put its weight behind the launch: Komdigi and Indosat presented the platform together in Jakarta on February 25, 2026. Minister Meutya Hafid framed the goal plainly: technology must reach everyone, without language or cultural barriers. Patrick Walujo, GoTo’s chief executive, called the 70B release a statement of support for local technology. Vikram Sinha, Indosat’s president director, was more specific about the infrastructure. “Sahabat-AI is a national asset,” he said. “GPU Merdeka stores and processes data directly at home.” The model speaks five Indonesian languages as first-class citizens: Bahasa Indonesia, Javanese, Sundanese, Balinese and Batak. Industry reviewers who tested the 70B model against global alternatives rate it the only option that handles Javanese well enough for production use. For a bank or a retailer whose customers write in those languages, that capability has no substitute at any price. Hosting inside Indonesia solves the compliance problem by construction. Prompts stay on Indonesian soil, processed by an Indonesian operator, under Indonesian law. The managed inference service on GPU Merdeka prices at roughly IDR 7,500 per million tokens, which industry comparisons put at parity with the cheapest tier of global API pricing. For most deployment sizes, compliance comes free with the subscription. The honest weakness is the ecosystem. The model family has accumulated about 35,000 downloads on Hugging Face, mostly of the smaller 8B and 9B variants. Independent review databases count fewer than 40 production deployments with public reviews. Tooling for fine-tuning, monitoring and evaluation is thin next to what surrounds the global models. A company choosing this path in 2026 buys compliance and language fit, and accepts that more of the integration work is self-service. It cannot yet buy the finished product it wants off a shelf, and whether a specific capability exists cannot be verified from a brochure; it has to be tested on the company’s own tasks. Path three: self-hosting the open weights The third path takes the same model and moves it in-house. Sahabat-AI ships open weights under permissive licenses, so a company with its own GPUs can run the model entirely on its own hardware. Nothing leaves the building. The economics flip at scale. Serving 50 million Bahasa tokens per day through a global API costs roughly 8,000 dollars per month. A self-hosted node with eight H100 GPUs costs roughly 12,000 dollars per month all-in, and then each additional token is free at the margin. Below a few hundred million tokens monthly, the API wins. Above it, ownership wins, and the gap widens with volume. A company processing millions of customer conversations can reach the crossover within its first year. The costs that decide this path are not the GPUs. Self-hosting requires MLOps capability that most Indonesian companies do not have on staff: capacity planning, model updates, evaluation pipelines, incident response. Hiring or training that team is a multi-quarter project. The public-sector and bank deployments that succeeded started 12 to 18 months before their deadlines. There is a middle version of this path that most SMEs should hear about. Managed inference on GPU Merdeka gives the data residency of self-hosting without the operations team. The token price is close to global API parity. For companies between the subscription tier and the hardware tier, this middle seat is usually the right answer. The three paths, side by side The table compresses the trade-offs. Prices are list or estimate levels as of late 2026 and change often. Path Monthly cost level Time to first output PDP Law position Best fit Global subscriptions 20-30 USD per user Days Transfers need a legal basis; keep personal data out Internal work, English output, small teams Sahabat-AI managed (GPU Merdeka) Near global API parity, around IDR 7,500 per million tokens Weeks Data stays in Indonesia Customer-facing Bahasa and local-language service Self-hosted 70B node About 12,000 USD for an 8x H100 node, marginal tokens free 2-4 quarters including team Fully on-premises High-volume processors with MLOps staff Ref: costs compiled from vendor list prices and industry deployment guides cited in the Sources section. The work after the chat Whichever path a company picks, the subscription is the cheap part. The expensive part is the same for all three: connecting the model to the work. This is what the e27 column on Southeast Asian SMEs argued in November 2026. Most small businesses in the region already run on WhatsApp, the author noted, so the first AI purchase should not be a chatbot. It should be the system that takes what the chatbot hears and does the follow-up work: matching payments, updating the order sheet, filing the receipt. That integration work is where the real money goes, and it is also where vendors’ price lists go quiet. An integration project for a mid-sized retailer, connecting an AI service to its order flow and accounting, typically costs more in implementation labor in the first year than in model fees. The labor is billed in rupiah, by local firms, which is one more reason the domestic ecosystem around the models matters as much as the models themselves. Indonesia’s largest AI operator learned the same lesson at national scale. Indosat used AI-driven personalization on its own network and cut user churn by 50 percent over 90 days while lifting average revenue per user more than 6 percent, to 2.62 dollars. The capital expenditure saved through AI-optimized network planning reached about 86.5 million dollars. Asked what he would change, chief executive Vikram Sinha did not point at the technology. He said he would have pushed adoption across the organization harder and earlier. The machines were never the bottleneck. The habits were. Who should buy which path The choice reduces to three questions: where does the data live, what languages do the customers speak, and how many tokens per month. A company under 50 staff, doing internal work, with English-capable staff: buy the global subscriptions, keep personal data out, spend the savings on an integration project for the two workflows that actually cost time. A company with customer-facing service in Bahasa or local languages, handling personal data, with no MLOps team: use the managed sovereign route, and treat the language coverage of Javanese, Sundanese, Balinese or Batak as the feature that justifies the switch. A company processing millions of tokens daily, with or willing to build an operations team: self-host the open weights, accept the 12,000 dollar monthly floor, and stop paying per token at scale. The two adoption numbers will not converge on their own. The 54 percent learned to use AI because it was free and fun. The 14 percent will climb only when the work after the chat gets built, and the three price tags above are what the building actually costs in Indonesia in 2026. Data basis Adoption figures come from the e-Conomy SEA 2025 report by Google, Temasek and Bain & Company, released in November 2025, and from Microsoft adoption tracking cited in the accompanying research. Enterprise figures for Claude and Lazada come from the tool operators’ published usage data. Sahabat-AI capability, pricing and launch details come from the operators’ announcements of February 25, 2026 and industry deployment guides published in 2026. Cost estimates for the three paths use list prices and the industry-standard deployment benchmarks cited in the text; all are estimates and change with pricing updates. The Personal Data Protection Law framework follows UU PDP 27/2022, GR 33/2026 and OJK guidance as described in published legal analyses. The Fore Kopi breach report is from November 2026. Data as of November 2026. Limitations The 54 percent and 14 percent figures measure different populations with different survey methods, so the 3.65 ratio is indicative rather than exact. Token and subscription prices shift quarterly, and the self-hosting crossover point depends on utilization that no public dataset discloses. Production deployment counts for Sahabat-AI cannot be verified beyond public reviews. Regulatory interpretation of GR 33/2026 is still settling in practice. Sources BusinessWorld — e-Conomy SEA 2025 regional GenAI adoption figures, October 2026 VOI — Google Indonesia launch remarks, 18M daily images, November 2025 ITBeat — Sahabat-AI platform launch with Komdigi, February 2026 Technologue — GoTo and Indosat executive remarks on Sahabat-AI 70B, 2026 ITNews Asia — Indosat AI Grid results and Sinha remarks, 2026 Software Listing — SEA sovereign LLM deployment costs and benchmarks, 2026 e27 — SME chatbot workflows column, November 2026 CNBC Indonesia — Fore Kopi data breach report to Komdigi, November 2026 Lexology — Indonesia AI legal framework snapshot, 2026 About Global Infinity Labs Global Infinity Labs is an AI innovation company specializing in MCP data services. This article was prepared by our expert team using insights and data provided through our MCP platforms. Contact: support@globalinfinitylabs.com
More
Korea Nakajo Accelerates Overseas Expansion with Commitment to Quality and Environmental Management SeaPRwire

Korea Nakajo Accelerates Overseas Expansion with Commitment to Quality and Environmental Management

Gimpo, Korea – November 16, 2026 – (SeaPRwire) – Korea Nakajo Co., Ltd. (CEO Jong-moon Choi), a manufacturer of commercial ice machines and shaved ice machines, is accelerating its expansion into overseas markets. Its push is built on quality and environmental management systems that meet international standards, along with its technical expertise. The company plans to become more competitive globally by strengthening production management, ensuring product safety, and preparing to meet each country’s environmental regulations. Korea Nakajo has drawn on its expertise in developing and manufacturing commercial ice machines and shaved ice machines to keep improving production efficiency and product quality. The company has built management systems based on international standards and is focused on improving quality across the entire product lifecycle, from development and production to after-sales service. As part of these efforts, Korea Nakajo has earned ISO 9001 (Quality Management System) and ISO 14001 (Environmental Management System) certifications. These international standards help companies maintain consistent product and service quality and manage the environmental impact of their operations. Building on these certifications, Korea Nakajo has standardized its production and quality control procedures, and its management system also takes environmental factors into account. The company is also obtaining international certifications and meeting regulatory requirements to address each country’s product safety, hygiene, and environmental standards. Korea Nakajo’s products are certified under CE, CB, and ETL for product safety, and the company is meeting NSF requirements for commercial ice machine hygiene and UKCA (UK Conformity Assessed) marking requirements for the United Kingdom. Korea Nakajo is also addressing European requirements such as food contact material (FCM) migration testing, WEEE, and POPs regulations, laying the groundwork for further overseas expansion. Korea Nakajo’s in-house R&D also supports its global competitiveness. The company holds INNOBIZ certification, which recognizes technology-driven small and medium-sized enterprises, and its ongoing research and development continues to improve ice-making performance and make hygiene maintenance easier. Several models include automatic cleaning and UV sterilization, making hygiene management more efficient and the machines easier to use. This combination of quality management and technology is driving the company’s overseas growth. Under its own brand, ICEVAN, Korea Nakajo exports to more than 20 countries across Europe, Asia, the Americas, and the Middle East. With its international certifications in place, the company is building partnerships with global distributors and business partners. “Strong product performance alone isn’t enough to compete in the global market. Quality and environmental management that meets international standards matters just as much,” said Jong-moon Choi, CEO of Korea Nakajo. “We will continue to innovate and strengthen our quality and environmental management systems to earn the trust of customers overseas and build ICEVAN’s position in the global ice machine market.” For more information about Korea Nakajo, please visit https://icevan.co.kr/inc_html/en/index_en.html. Media Contact Company: Korea Nakajo Co., Ltd. Contact: Sales Representative Email: nakajo@icevan.co.kr Website: https://icevan.co.kr/inc_html/en/index_en.html
More
Family office rule churn: the compliance machine outsiders never see SeaPRwire

Family office rule churn: the compliance machine outsiders never see

Singapore rewrote its family office framework in June 2026 while London counts departures. The people inside the paperwork carry the cost. In the days after the British budget that abolished the non-dom regime, a tax lawyer in London got a text message from a client. “You need to drive me to the airport,” the message read. The client was worth about 200 million pounds. Another client had already planned for the worst case and boarded a plane back to South Africa the same day. James Quarmby, founding partner at Stephenson Harwood, told City A.M. that “dozens and dozens” of his clients would now leave, and that replicated across the City the number runs into the thousands. Six months later and half a world away, a tax director in Singapore named Shaun Zheng was answering a different question from a different set of wealthy families. Singapore’s regulator had just rewritten the rulebook under which single family offices operate. His clients, he told the Business Times, were asking “whether we still qualify” and “what extra work does this create for us.” His team had hoped existing family offices would be grandfathered in. They were not. Two jurisdictions, two very different policy moves, one shared experience. The families move, or re-paper, or wait. In between stands a layer of professionals, lawyers, tax advisers, external asset managers, compliance officers, whose year is consumed by other people’s rule changes. This piece is about that layer: how much of it there is, what its members say in public, and why the cost they absorb never shows up in any policy scorecard. How many people live inside this machine The population is larger than the phrase “family office” suggests. Singapore’s central bank confirmed in a 2026 parliamentary reply that more than 2,000 single family offices held tax incentives by the end of 2025, while the city’s assets under management grew 10.1 percent to 6.7 trillion Singapore dollars. Industry estimates put the global count of family offices at around 10,000, and estimates vary because no registry exists; a firm with two employees advising one family and a firm with eighty managing five billion dollars both count. The UK side of the machine is processing departures. HMRC figures published in July 2026 show about 9,000 non-doms left in the 2024-25 tax year, with more than a thousand departing before the reforms formally took effect. The Office for Budget Responsibility’s central scenario anticipates 10,800 non-dom departures a year, with a downside case above 12,000. On the survey side, the Saltus Wealth Index found 30 percent of UK high net worth individuals would consider leaving the country, including 8 percent actively exploring a move within twelve months. Among households worth at least 5 million pounds, the actively-considering share reaches 15 percent. These numbers do not describe the same thing, and one of the year’s sharpest data conflicts sits right here. New World Wealth’s migration estimates put Britain’s 2025 millionaire net outflow at 16,500, the largest of any country. The OBR’s own modeling of the non-dom changes implies roughly 1,200 departures in the relevant base. Both cannot be right about the same people; they measure different populations with different definitions, and this article presents both rather than picking one. What is not in dispute is the direction of adviser workload, which every professional quoted below describes as up. The client base the machine serves keeps growing, which is why the workload compounds instead of fading. Knight Frank’s The Wealth Report 2026 counts 713,626 ultra-high-net-worth individuals worldwide, up from 551,435 in 2021, and projects 948,241 by 2031. In the 2021-2026 window that growth ran at 89 net new UHNW individuals per day; the next five years are forecast at 129 per day. The United States holds 251,352 of them, about 35 percent of the world total, and is forecast to reach 41 percent by 2031. Jurisdiction UHNW individuals 2026 Growth to 2031 United States 251,352 +54 percent United Kingdom 27,876 +11 percent Saudi Arabia 4,388 +63 percent UAE 4,851 +36 percent Source: Knight Frank, The Wealth Report 2026 , April 2026. The growth league table is the competitive map. Saudi Arabia’s 63 percent projected growth and the UAE’s 36 percent sit far above the UK’s 11 percent. Dubai and Singapore are often framed as rivals for the same families, but advisory commentary through 2026 increasingly describes them as complementary: Singapore anchoring Southeast Asian deal flow and family structures, Dubai anchoring Gulf capital and, increasingly, the departing British. The clients are mobile. The rules that decide where they land are not, and that mismatch is the business. What the professionals say in public Read the interviews together and the vocabulary repeats: deadlines, re-papering, queues, and questions about who still qualifies. In Singapore, the revised regime that took effect on June 15, 2026 replaced case-by-case approval with a notification model built on three requirements: notify the regulator, maintain a relationship with a regulated bank, and file an annual return. Existing offices have one year, until June 2027, to align. Zheng, director of tax services at CLA Global TS, wanted grandfathering and did not get it. His clients now want formal confirmation of eligibility, funding source rules and reporting processes. David Ng, co-founder of Arki Finance, said some offices need more time to check whether trusts, foundations and charitable vehicles fall within the qualifying criteria, and that multigenerational setups need the most review. Christos Anagnostopoulos, head of family office solutions for Asia at Julius Baer, noted that many offices are operationally lean and face real constraints building the internal capability the new filing obligations demand. Anuj Kagalwala, PwC’s Asia-Pacific private wealth leader, reported more queries arriving in two waves, one now and a second near the deadline. The competitive effects drew the sharpest comment. Lester Tay, chief executive of Rockstead Capital, a Singapore multi-family office managing 1.8 billion Singapore dollars, said the new rules tilt the field toward private banks. Because every qualifying office must bank with a MAS-licensed institution, the banks gain a direct channel to family office clients, and boutique external asset managers get squeezed toward the sub-20-million-dollar segment. His own workaround: the S$5 million non-SFO scheme that does not require a private banking account. Not everyone reads the change as burden. Wyn James of IQ-EQ called the revised regime genuinely pragmatic: low friction at entry, with enough oversight to keep credibility. Ryan Lin, director of Bayfront Law, said the notification model provides greater certainty and reduces onboarding friction while preserving anti-money-laundering safeguards. Both frame the rule change as a product improvement, not a tax. Underneath the compliance commentary sits a capability gap that makes every rule change hit harder. A Lombard Odier survey of more than 390 high-net-worth individuals across Asia-Pacific, published in June 2026, found that fewer than a third seek professional advice on succession planning and inheritance structuring, and only about 13 percent have sought advice on family governance and conflict resolution. Lester Tay sees the consequence daily: founders who treat legacy planning as an afterthought, and whose investment decision-making has no documented successor. When a regulator tightens eligibility, the families with the thinnest internal records have the most rebuilding to do, and they discover the gap at the worst possible moment, under a deadline. The London commentary is blunter. Marc Acheson, global wealth specialist at Utmost, said the UK is “driving wealth creators away” and described a second wave of departures now including long-term nationals and business owners. Daniel Lewin, tax partner at Katten, called the Sunday Times Rich List “painful reading” and noted that one in six names from two years earlier had dropped off. Daniel Channing of Whitmill described wealthy individuals forming a “queue” to leave, delayed by businesses, schooling and personal affairs. Matthew Gamman of Cone Marshall captured the dynamic that keeps the queue moving: “The threat of closing the door does not necessarily keep people in the room, it may simply persuade them to leave before it shuts.” The counter-voices exist here too. Iain Tait, head of the private investment office at London and Capital, called the extension of the temporary repatriation facility “welcome news” even while calling the overall impact significant. Richard Jordan of Trowers & Hamlins argued the fiscal payoff will disappoint by 2027 and that steady tinkering would have beat wholesale abolition. Their position is not that nothing happened, but that the magnitude is contested and the story is not one-way. The 2026 Sunday Times Rich List gave the departure story its most quotable evidence. The compiler, Robert Watts, recorded the billionaire count falling to 157 and found that one in six names from the list of two years earlier had vanished, alongside a sharp rise in British nationals now resident in Dubai, Switzerland and Monaco. Eight UAE-based figures now appear among the 350 richest. Henley & Partners’ private wealth migration report, using New World Wealth data, ranks Britain as the largest exporter of high-net-worth individuals among major economies, with the UAE, Italy and Switzerland as principal beneficiaries. Each list and estimate uses a different definition of who left and when, which is precisely why the OBR’s 1,200 and New World Wealth’s 16,500 can coexist without either being a simple error. Two forces, not one The obvious reading is that tax policy causes all of this. The evidence supports a second condition that has to be true at the same time: the jurisdictions competing for mobile wealth now differentiate themselves by changing rules, frequently, in public, and on deadline. Singapore’s June 2026 overhaul is the cleaner example. The city-state had already doubled the asset threshold for the main tax incentive to 50 million Singapore dollars, required three investment professionals, and tightened source-of-wealth review after the 2023 money laundering case that exposed 3 billion Singapore dollars in assets. Then, within weeks, it relaxed parts of the economic substance test, allowing a 13O office to start with one investment professional, while expanding funding-source scrutiny. The May 2026 Financial Action Task Force rating, the highest monitoring tier, was used in official commentary as proof the tightening worked. Each move repositions Singapore against rivals: Hong Kong, the UAE, and newer entrants. Advisers quoted in the Singapore press describe exactly that contest, with one noting that sophisticated families now run multijurisdiction models across the region rather than a single hub. The UK moved in the opposite direction and produced the same professional workload. Abolition of the non-dom regime from April 2025, a four-year foreign income window for new arrivals, inheritance tax extended to worldwide assets after ten years of residence, and a ten-year tail after departure. Each element forced existing setups to be unwound or rebuilt. Whether the policy raises or loses revenue is contested between the OBR’s modeling and the estimates of the migration research firms. What is not contested is that every element generated advisory work. The two conditions compound. Rule churn alone would be manageable if rules converged; every major wealth jurisdiction converging on the same standard would mean one re-papering. Jurisdictional competition alone would be manageable if rules were stable; families would pick a home and stay. Together they guarantee that the compliance layer re-checks every setup repeatedly, against moving targets, on client money. The machine outsiders never see Ask who pays for this and the honest answer is: nobody planned to. The family pays adviser fees, but the quote the family sees is the setup fee, not the re-papering fee triggered by the next rule change. The regulator sees compliance rates, filing volumes and enforcement statistics, not the number of professional hours consumed aligning a legitimate holding with a revised form. The adviser firms book the hours as revenue, which is why you will not hear them complain too loudly. Three costs fall through the gap: The duplicated review: the same family holding gets re-reviewed by lawyers in two or three jurisdictions after every relevant rule change, and no consolidated figure for that spend is published anywhere. The Monetary Authority of Singapore does not disclose an estimate of industry compliance cost for the rule change, and HMRC’s impact assessments for the non-dom reforms model taxpayer behavior, not adviser workload. The timing cost: families defer real decisions, succession, asset sales, relocations, until rules settle, and deferral has a price that no filing captures. The error cost at the margin: advisers describe clients attempting to self-assess eligibility under the new notification model, and a wrong self-assessment surfaces only at audit. The professionals themselves are the shock absorbers, and the market is reorganizing around them. Multi-family offices report demand from single-family offices that cannot staff the new obligations. External asset managers target the sub-threshold segment that the new rules push out of the tax-incentive system. Compliance hiring in Singapore wealth management runs hot. The machine does not shrink after each rule change; it digests the change and grows. None of the public commentary disputes the policies’ stated goals: cleaner money, fairer taxation, credible centers. The professionals quoted here mostly accept those goals. Their testimony is about something narrower and easier to overlook. In a market where jurisdictions compete by moving rules, the movement itself is a cost line, and it lands on a layer of the industry that both the clients and the policymakers treat as overhead. The 200-million-pound client who texted his lawyer from the airport saw a tax decision. The lawyer saw a quarter of unwound work, and a pipeline of it behind. Sourcing note This piece rests on named-public-figure interviews and regulatory documents published between March 2025 and November 2026. Singapore quotes come from Business Times syndicated reporting on the June 2026 overhaul, Asian Private Banker interviews, and Chinese-language coverage by Economic Daily News and Fortune Times. UK quotes come from City A.M. budget-night reporting, trade press statements on the Sunday Times Rich List, and the Saltus Wealth Index and Spear’s tax survey summaries of October 2026. Official figures come from MAS parliamentary replies, HMRC publications and OBR scenarios. Where estimates conflict, chiefly New World Wealth’s 16,500 millionaire outflow against the OBR’s roughly 1,200 non-dom departures, both figures are presented with their sources rather than reconciled. Several Singapore trade publications are secondary sources with limited editorial history; their factual claims were cross-checked against regulator statements where possible. Data as of November 2026. Ethics note All quotations are drawn from named individuals speaking in published interviews or press statements, and every attribution names the person and publication. No anonymous sources were used. The piece describes professional workload and client behavior; it makes no allegation of wrongdoing against any family, adviser, firm or regulator, and the one criminal matter referenced, Singapore’s 2023 money laundering case, is cited only as public record of the policy context. Jurisdiction comparisons describe policy outcomes, not recommendations. Sources Daily8 — Existing SFOs evaluate revised MAS rules, Business Times syndication, June 2026 Syntho — MAS rules one-year countdown, adviser remarks, June 2026 Asian Private Banker — Rockstead Capital chief executive interview, 2026 Fortune Times — RWM director Walter Ng on Singapore’s tightening, August 2026 (in Chinese) Economic Daily News — Singapore lawyers on compliance thresholds, March 2026 (in Chinese) MAS — Parliamentary reply on family offices, August 2026 Wealth Investment News — Utmost and Katten remarks on UK Rich List, 2026 City A.M. — Tax advisers on non-dom abolition, budget-night reporting, 2024 SITNAS — Saltus Wealth Index and Spear’s tax survey summaries, October 2026 deVere Group — HMRC non-dom departure figures and OBR scenarios, July 2026 Knight Frank — The Wealth Report 2026, UHNWI populations and forecasts, April 2026 BCG via Chinanews — Singapore wealth management growth and SFO count, May 2026 (in Chinese) Global SWF — Gulf sovereign investor deployment, 2026 annual report World Inequality Report — Global wealth concentration, 2026 Mirabello Consultancy — UK tax reform timeline summary, 2026 Chinanews — BCG and Private Client rankings conflict on Asia’s top wealth hub, 2026 About Global Infinity Labs Global Infinity Labs is an AI innovation company specializing in MCP data services. This article was prepared by our expert team using insights and data provided through our MCP platforms. Contact: support@globalinfinitylabs.com
More
The new-home glut is here. The resale glut never came SeaPRwire

The new-home glut is here. The resale glut never came

New-home supply sits at 8.4 months while resale holds at 4.6 in late 2026. What the split market means for builders, sellers and buyers through 2027. In September 2026 the median new single-family house in the United States sold for 400,400 dollars. The median existing house sold for 434,100 dollars two months earlier. A brand new home now costs less than a used one. That inversion is the clearest signal this housing cycle has produced, and it says something specific: the two halves of the housing market have stopped behaving like one market. Builders have too much inventory. Homeowners with mortgages have too little willingness to sell. Both conditions come from the same cause, a decade of cheap loans followed by a decade of expensive ones, but they produce opposite problems. This article walks through the split, the two forces that keep it in place, and the two numbers that will show which side cracks first. One market, two inventories The Census Bureau’s new residential sales report for September 2026 shows the stress on the builder side. New home sales ran at a 676,000 annual rate, down 5.3 percent from a year earlier. There were 476,000 new homes for sale, and at the current sales pace that inventory represents 8.4 months of supply. Anything above 6 months is conventionally read as a buyer’s market. The resale side, tracked by the National Association of Realtors, sits on the other side of that line: existing home inventory equals 4.6 months of supply, with the median price up 2.0 percent year over year at 434,100 dollars. Measure New homes Existing homes Sales pace (annualized) 676,000 4,060,000 Change from a year earlier -5.3 percent Near flat Median price 400,400 USD 434,100 USD Months of supply 8.4 4.6 Homes for sale 476,000 Tight, locked-in owners Sources: US Census Bureau new residential sales, September 2026 and NAR research and statistics . Prices confirm the same split from a different angle. The FHFA House Price Index rose 0.3 percent month over month and 2.4 percent year over year through August 2026, reaching an index level of 444.6. Appreciation is positive but slow, and it hides the divergence underneath. Regional trackers through 2026 show New York up 5.1 percent and Chicago up 5.7 percent while San Francisco gained 0.5 percent and Atlanta sat flat. Supply-constrained coastal and Midwest markets keep climbing. Sun Belt markets where builders concentrated their pipelines are coming back to earth. Market Price change, latest reading Chicago +5.7 percent New York +5.1 percent San Francisco +0.5 percent Atlanta Flat Source: regional index tracking compiled by Lighthouse Macro , 2026. So the headline market is frozen, and the frozen surface conceals one submarket with a glut and one with a drought. The interesting question is what holds the split in place. Force one: the mortgage lock-in The first force is arithmetic. Roughly 80 percent of outstanding American mortgages carry rates of 6 percent or lower, a legacy of the pandemic-era refinancing wave. The average 30-year fixed rate stood at 6.67 percent in August 2026, per the Freddie Mac survey, and the cost of borrowing against home equity has been climbing: the national average HELOC rate reached 7.49 percent in early November. An owner holding a 3 percent mortgage who sells into this market replaces a 3 percent loan with a 7 percent one. On the median home that difference adds several hundred dollars to the monthly payment for the same house. Most owners respond by staying. The result is a resale market that cannot clear, not because nobody wants to buy, but because almost nobody wants to be the seller. This is why the resale glut never came. In a normal downturn, falling demand forces price cuts, and price cuts clear inventory. Here the correction mechanism is disconnected. Existing inventory stays scarce by construction, because the stock of cheap mortgages is scarce. Prices on the resale side hold through scarcity rather than through demand, and buyers compete over a shrinking selection of used homes. Rate forecasts say the lock does not open soon. Fannie Mae’s 2026 forecast puts the 30-year fixed at an average of 6.3 percent for this year and 6.2 percent for 2027. Economists who model the market note that a sustained move below 6 percent would be needed to bring meaningful listing volume back. Nothing in the current rate path approaches that level. Estimates of how many owners would list at various rate levels vary, and no public dataset does disclose the true distribution of owner rate sensitivity, but the direction is not in doubt: every tenth of a point of rate relief brings some sellers to market, and there has been very little relief. One workaround niche has grown around the lock. About 12.2 million active mortgages, some 23 percent of the total, are assumable: a buyer can take over the seller’s loan and its rate. Assumption deals need lender approval and big down payments to cover the equity gap between the old loan balance and the current price, so they remain a slice of the market rather than a solution. Their existence is itself evidence of the problem. Buyers will pay a premium just to inherit a mortgage rate. Force two: the builder incentive machine The second force is behavior. Homebuilders, unlike homeowners, can manufacture concessions. They own the land, control the product, and can subsidize a buyer’s interest rate directly through mortgage rate buydowns. The National Association of Home Builders survey data through 2026 shows an industry buying volume with margin. Builder sentiment has run below the 50 breakeven line for more than 21 consecutive months, with the Housing Market Index between 35 and 38 for most of the year. In August 2026, 35 percent of builders reported cutting prices, at an average reduction of 6 percent, and 63 percent offered sales incentives such as rate buydowns, closing cost help or upgrades. Earlier in the year the price-cutting share hit 40 percent, the highest since May 2020. Builder survey measure Early 2026 August 2026 Housing Market Index 37 About 35-38 Share cutting prices 40 percent 35 percent Average price cut 6 percent 6 percent Share offering incentives 65 percent 63 percent Source: NAHB survey data via industry summaries . This machine explains the price inversion. The 400,400 dollar median new home is not cheaper because construction got cheaper. Building materials producer prices rose 3.4 percent year over year through February 2026. It is cheaper because the product changed and the financing is subsidized. Median new-home square footage has declined from its 2015 peak, lots have shrunk, and construction has concentrated in lower-cost Southern markets with lighter land regulation. The buyer of a new home in 2026 gets a smaller house on a smaller lot farther out, with the monthly payment bought down at the builder’s expense. The used home on the other side of the comparison carries costs the median price hides. The median owner-occupied house in America is 42 years old, up from 31 years in 2005, and 47 percent were built before 1980. Homes of that age carry pre-modern wiring, galvanized plumbing, and roofs or heating systems at or past their service life. The age map follows the zoning map. New York’s median owner-occupied home is 64 years old, and Massachusetts and Rhode Island sit at 59. Nevada’s is 25 and Texas 28. Markets that blocked construction for decades are now selling the oldest stock at the highest prices, and buyers pay the scarcity premium knowing the repair bill comes with the house. From 2020 to 2024 the country added about 3.6 million owner-occupied homes, just 4 percent of the total stock, and the share of owner-occupied homes less than 14 years old fell from 18 percent in 2014 to 13 percent in 2024. The incentive machine works because builders answer to volume targets and public shareholders, not to a memory of their own mortgage. It also has a limit: margin. Large public builders like D.R. Horton and Lennar can absorb years of compression. Smaller private builders cannot, which is why consolidation accelerates in this phase of every cycle. History gives the scale of the current stress a ceiling. New-home months of supply peaked above 12 months in the 2008 crash, against 8.4 today. Overall months of supply in that downturn ran roughly double the current level. The NAHB index troughed at 8 in that cycle against 35 to 38 now. Today’s builder market is weak by the standard of the last decade and nowhere near distressed by the standard of 2008. The distinction matters for anyone reading the inventory numbers as a crash signal. Where the two forces are heading Both forces are already resolving, slowly, on the builder side. Housing starts fell to 1.246 million units annualized in October 2025, the lowest since June 2020, with single-family starts at 874,000. Single-family permits, the pipeline measure that leads starts by one to two months, ran at 878,000, well below the roughly 1 million pace needed to match household formation. Builders have stopped replenishing the glut at the source. The months of supply number will drift down over the next year as sales burn through a shrinking pipeline, unless sales fall faster than starts. Freddie Mac’s long-run estimate puts the national housing shortfall at 3 to 4 million units, and estimates of the true gap vary by methodology. The shortfall is the floor under national prices. It does not protect builders from their own inventory, because the shortfall is concentrated in markets where builders are not currently building. Rates are the other variable, and the forecast says stasis. With the 30-year fixed projected to average 6.2 to 6.3 percent through 2027, the lock-in holds. The resale shortage persists. The payment math that freezes sellers also freezes the arbitrage that would otherwise close the price inversion. New homes stay cheap relative to used ones not because the market decided new is better, but because only the new side is allowed to cut. Demand, meanwhile, has not died. Pending home sales rose 1.4 percent month over month and 3.2 percent year over year in April 2026, and existing sales held near a 4.0 million annual rate through the year, close to the lowest levels since the financial crisis but stable. The buyers still in the market are payment-sensitive and selective. They respond to the monthly payment, which is exactly the number builders can manipulate and sellers cannot. That asymmetry, more than any preference for new construction, explains why new homes kept selling while resales sat. Who gets hit first if demand weakens again The exposure ranking follows the margin math, and it runs in a predictable order. Small private builders absorb the first hit. They lack the balance sheets to fund rate buydowns at current scale, and their land positions are smaller and more leveraged. Consolidation accelerates: land and half-finished communities transfer to public builders at discounts. Land sellers and developers in exurban submarkets take the second hit. The lots already entitled for construction in the Sun Belt corridor lose value fastest when starts decline, because the discount is where the pipeline was thickest. Move-up resale sellers are third. They compete directly against subsidized new builds in the suburbs, and their homes lack the two selling points the builder offers: a new roof and a bought-down payment. In states that reassess on sale, the penalty compounds. A buyer in Florida or California who purchases an existing home takes on property taxes reset to full market value, while the seller’s capped assessment dies with the deed. The same house costs a different amount to own depending on when you bought it, which pushes marginal move-up buyers toward new construction in low-tax counties. Submarkets with the oldest housing stock are insulated the longest. Nearly half of American owner-occupied homes were built before 1980, and the median owner-occupied home is 42 years old. Used homes in supply-starved metros do not compete on price with a glut three states away. The renter side diverges from the builder story. The multifamily construction boom of the pandemic years is unwinding, with multifamily starts collapsed to 372,000 annualized in late 2025, and completions still feeding the rental market. Renters in newly delivered buildings see concessions. Renters in old buildings in strict zoning markets do not. The affordability problem resolves only through the slow channel. Nominal price appreciation has run at 1 to 2 percent while wages grow near 4 percent, so the gap between what the median household earns and what the median home costs closes by roughly 2 to 3 percentage points a year. At that pace the market normalizes over years, not quarters, and only in the segments where prices actually stagnate. Real home prices, adjusted for inflation, are flat to slightly negative, which is the healthy version of a correction: no crash, no relief either. The two numbers that will call the turn Two leading indicators cover this market, and they are independent of each other in method and source. The first is new-home months of supply, published monthly by the Census Bureau from hard inventory counts. The second is the NAHB share of builders cutting prices, from a monthly survey of builder behavior. One counts houses. The other asks businessmen what they are doing with prices. A deterioration confirmed by both is real; a divergence between them is usually a timing artifact. The call embedded in this analysis is that the split persists through 2027: builder inventory burns down slowly, resale stays locked, and national prices muddle through at low single digits. That call is wrong in four observable ways. New-home months of supply falls below 6 while sales rebound above 700,000 annualized. That means the glut cleared faster than the pipeline could shrink, and builders regain pricing power sooner than expected. Existing-home months of supply rises above 6. That means the lock-in broke, most likely through a rate move below 6 percent, and resale inventory floods the market. The price inversion would then close from the other direction, with used homes getting cheaper. The NAHB price-cut share climbs back above 40 percent while the Housing Market Index breaks below 32. That means the incentive machine hit its margin floor, and volume targets are being abandoned. Watch for outright project cancellations in that case. The 30-year fixed rate sustains below 6 percent for a quarter. Every forecast in this article is conditioned on rates staying above that line, and the whole split-market arrangement reorganizes if the condition fails. The split market is not a contradiction. It is what a housing cycle looks like when the price of moving is a mortgage reset, and only one side of the market is allowed to negotiate. The inversion in the medians will look strange until rates break 6 percent or the last of the cheap loans rolls off. Neither happens soon. Data basis New-home figures come from the US Census Bureau new residential sales report for September 2026, released in late October 2026. Resale figures come from National Association of Realtors existing-home sales reporting for July 2026. The FHFA House Price Index covers August 2026. Builder survey data comes from NAHB Housing Market Index releases through 2026. Mortgage rate levels come from the Freddie Mac Primary Mortgage Market Survey and Bankrate lender surveys. The rate forecast is Fannie Mae’s published 2026 housing outlook. Housing stock age and regional price divergence figures come from the 2024 American Community Survey and index trackers cited in the text. Data as of November 2026. Limitations The Census and NAR figures use different sampling frames and revision schedules, so month-to-month comparisons across the two sides carry noise. The 80 percent lock-in share is an industry estimate, and estimates vary by whether assumptions and assumable loans are included. Builder survey shares are self-reported. The insulation of old-housing-stock metros is an inference from price divergence, not a tested model, and a recession would stress every segment at once. Sources US Census Bureau — New residential sales, September 2026, released October 2026 FHFA — House Price Index, August 2026 data NAR — Existing-home sales research and statistics, 2026 EffectiveAgents — NAHB builder incentive and price-cut survey summaries, 2026 KL Brown & Associates — Raymond James commentary with Fannie Mae rate forecast and lock-in analysis, May 2026 Lighthouse Macro — Housing starts, permits, builder sentiment and regional price divergence, 2026 Bankrate — HELOC rate survey, November 2026 About Global Infinity Labs Global Infinity Labs is an AI innovation company specializing in MCP data services. This article was prepared by our expert team using insights and data provided through our MCP platforms. Contact: support@globalinfinitylabs.com
More
Indonesia’s AI gap: 54 percent use it, 14 percent work with it SeaPRwire

Indonesia’s AI gap: 54 percent use it, 14 percent work with it

The numbers say everyone touches AI tools in 2026. The work says few have wired them into the business. Three paths for Indonesian companies, three price tags. An owner of a food business in Jakarta runs the whole company on WhatsApp. Orders arrive there. Suppliers send invoices there. Staff photograph receipts and send them to the same group chat. When someone offers to sell her “AI”, the pitch is usually a chatbot that answers customers faster. That pitch misses the job. The job is the work after the chat: matching invoices to orders, chasing payments, filing records. This gap between talking to AI and working with AI shows up in the national numbers. Indonesia has one of the highest rates of casual AI use in Southeast Asia. It has one of the lowest rates of AI inside business workflows. The distance between those two numbers is where most companies are stuck, and the three ways out of that position cost very different amounts of money and time. The two numbers measure different people The e-Conomy SEA 2025 report, published by Google, Temasek and Bain & Company, puts standalone generative AI use among Indonesia’s working population at 54 percent. The six-country regional average is 58 percent. Thailand sits at 63 percent, Vietnam at 65 percent, Malaysia at 68 percent and Singapore at 70 percent. Indonesia’s internet penetration is 81 percent, so the base is there. The consumer-side numbers run even hotter. The same report says 79 percent of Indonesian users have taught themselves to use AI tools. Veronica Utami, Google’s country director in Indonesia, said at the report launch that 80 percent of users in the country interact with AI daily. She gave one detail that shows the scale: Indonesians generate 18 million images per day on the Nano Banana image feature alone. Revenue from AI-enabled apps in Indonesia grew 127 percent year on year, the fastest in the region. Consumer interest in AI runs about three times the global average, by her account. The enterprise-side numbers tell the other half of the story. Microsoft’s own adoption tracking puts AI use inside Indonesian organizations at 14.1 percent in one measurement period and 14.8 percent in the next. Compare that with the 54 percent individual figure and the ratio is about 3.65 to 1. People play with AI at home. Companies have not rebuilt their work around it. The pattern holds across tools. Claude’s usage tracking shows 23 percent penetration in Indonesian computer and math occupations and 25 percent in education. Lazada reports 31 percent of its Indonesian sellers using its AI sales assistant. These are real numbers, and they are all in the 20 to 30 percent band, far above the 14 percent organization figure and far below the 54 percent consumer figure. The regional spread shows Indonesia is not an outlier in the casual-use number. It is an outlier in the size of the gap. Market Working-population GenAI use Internet penetration Singapore 70 percent Above 90 percent Malaysia 68 percent Above 90 percent Vietnam 65 percent 87 percent Thailand 63 percent Above 90 percent Indonesia 54 percent 81 percent Philippines 53 percent 70 percent Source: Google, Temasek and Bain e-Conomy SEA 2025 , as reported October 2026. Where AI actually touched Indonesian commerce in 2025 was not the office. It was the selling floor. Video commerce, the format where sellers stream and sell inside chat-adjacent apps, grew its active seller count 75 percent year on year to 800,000 sellers. Transaction volume rose 90 percent to 2.6 billion transactions. The overall digital economy is closing on 100 billion dollars in gross merchandise value, with e-commerce at 71 billion dollars. Those transactions increasingly ride on AI-adjacent rails: recommendation feeds, AI-dubbed product videos, automated customer replies. The consumer adoption number and the commerce number describe the same people. The surveys behind these numbers measure different populations, so the exact gap between 54 and 14 percent is an estimate rather than a measurement. Estimates vary on where casual use ends and operational use begins. The direction does not. Casual use runs far ahead of operational use, and the gap is wider in Indonesia than in most of the region because the consumer base grew so fast. Why the obvious answer stalls A company that wants to close this gap usually starts the same way. Someone buys subscriptions to a global AI tool, hands them out, and waits for productivity. In Indonesia this stalls for three reasons that stack on top of each other. The first is language. Global models speak English first. Bahasa Indonesia works, but the languages of daily commerce do not. A shop owner in Surabaya who answers customers in Javanese or Sundanese cannot hand those conversations to a model that fumbles both. The e-Conomy data shows consumer AI activity concentrated in exactly these local-language interactions, which is why adoption exploded among individuals and crawled inside firms. The second is money in the wrong currency. Global AI subscriptions bill in US dollars, typically 20 to 30 dollars per user per month. For a 50-person company that is 12,000 to 18,000 dollars a year before any integration work. That is not a large number by global standards. Against Indonesian SME margins it competes directly with payroll. The third reason is legal, and it arrived on a schedule. Indonesia’s Personal Data Protection Law, UU PDP 27/2022, took full effect in October 2024. Its implementing regulation GR 33/2026 requires controllers to keep a register of processing activities, run impact assessments before higher-risk processing, and honor the right to object to automated decisions. Cross-border transfers of personal data need a legal basis, and remote access from outside Indonesia counts. The Otoritas Jasa Keuangan has issued AI governance guidance for banks, and the regulator treats customer data inside AI prompts as personal data like any other. Enforcement is no longer theoretical. In November 2026, the coffee chain Fore Kopi reported a leak of 5.7 million customer accounts to Komdigi, the communications ministry. The company filed the report itself and apologized publicly. Whatever the investigation finds, the case lands in a system that now has a working complaint channel and a 72-hour breach notification rule for controllers. None of this makes global AI tools illegal. It makes them a compliance project. Sending customer chat histories through a model hosted in another country now needs a documented lawful basis, a transfer mechanism and an impact assessment. For a bank, the review alone can take six months. For a two-person trading company, it takes either a lawyer they cannot afford or a shortcut they should not take. Path one: global subscriptions, kept in their lane The cheapest path is also the most common. A company buys seats on a global tool and uses it where the compliance surface is thin: internal English-language work, drafting, summarizing, coding, analyzing its own non-personal data. The cost is known and small. A team seat runs 25 to 30 dollars per user per month at current list prices. Deployment takes an afternoon. The tools are the best in the world at general reasoning and English output. For a company whose staff already read and write English at work, this path delivers value in week one. The lane restrictions are the whole point. Customer data stays out. Personal data stays out. Anything the PDP Law would call processing of personal data stays out, because the models sit on servers in other countries and no company wants to build a cross-border transfer file for a chatbot subscription. One Jakarta bank described in a 2026 industry guide kept GPT-4o for internal English tooling while moving its customer-facing Bahasa chatbot to local infrastructure. That split is the pattern the market found: global tools for the back office, local systems for the front line. The weakness of this path is that it barely touches the 14 percent number. Internal drafting makes individuals faster. It does not automate invoices, reconcile orders or answer Javanese customers. Companies that stop here stay stuck between the two adoption curves, paying a new subscription and reporting no change in output. Path two: the sovereign hosted route The second path is new in 2026. Sahabat-AI is the open-source model family built by GoTo and Indosat Ooredoo Hutchison, now at 70 billion parameters, running on Lintasarta’s GPU Merdeka infrastructure inside Indonesia. The project had a public origin worth recalling: NVIDIA’s chief executive introduced it at Indonesia AI Day in 2024, the model family reached 70 billion parameters by mid-2025, and the consumer platform arrived in early 2026. The public chat service is free, on the web and inside the GoPay app. The government put its weight behind the launch: Komdigi and Indosat presented the platform together in Jakarta on February 25, 2026. Minister Meutya Hafid framed the goal plainly: technology must reach everyone, without language or cultural barriers. Patrick Walujo, GoTo’s chief executive, called the 70B release a statement of support for local technology. Vikram Sinha, Indosat’s president director, was more specific about the infrastructure. “Sahabat-AI is a national asset,” he said. “GPU Merdeka stores and processes data directly at home.” The model speaks five Indonesian languages as first-class citizens: Bahasa Indonesia, Javanese, Sundanese, Balinese and Batak. Industry reviewers who tested the 70B model against global alternatives rate it the only option that handles Javanese well enough for production use. For a bank or a retailer whose customers write in those languages, that capability has no substitute at any price. Hosting inside Indonesia solves the compliance problem by construction. Prompts stay on Indonesian soil, processed by an Indonesian operator, under Indonesian law. The managed inference service on GPU Merdeka prices at roughly IDR 7,500 per million tokens, which industry comparisons put at parity with the cheapest tier of global API pricing. For most deployment sizes, compliance comes free with the subscription. The honest weakness is the ecosystem. The model family has accumulated about 35,000 downloads on Hugging Face, mostly of the smaller 8B and 9B variants. Independent review databases count fewer than 40 production deployments with public reviews. Tooling for fine-tuning, monitoring and evaluation is thin next to what surrounds the global models. A company choosing this path in 2026 buys compliance and language fit, and accepts that more of the integration work is self-service. It cannot yet buy the finished product it wants off a shelf, and whether a specific capability exists cannot be verified from a brochure; it has to be tested on the company’s own tasks. Path three: self-hosting the open weights The third path takes the same model and moves it in-house. Sahabat-AI ships open weights under permissive licenses, so a company with its own GPUs can run the model entirely on its own hardware. Nothing leaves the building. The economics flip at scale. Serving 50 million Bahasa tokens per day through a global API costs roughly 8,000 dollars per month. A self-hosted node with eight H100 GPUs costs roughly 12,000 dollars per month all-in, and then each additional token is free at the margin. Below a few hundred million tokens monthly, the API wins. Above it, ownership wins, and the gap widens with volume. A company processing millions of customer conversations can reach the crossover within its first year. The costs that decide this path are not the GPUs. Self-hosting requires MLOps capability that most Indonesian companies do not have on staff: capacity planning, model updates, evaluation pipelines, incident response. Hiring or training that team is a multi-quarter project. The public-sector and bank deployments that succeeded started 12 to 18 months before their deadlines. There is a middle version of this path that most SMEs should hear about. Managed inference on GPU Merdeka gives the data residency of self-hosting without the operations team. The token price is close to global API parity. For companies between the subscription tier and the hardware tier, this middle seat is usually the right answer. The three paths, side by side The table compresses the trade-offs. Prices are list or estimate levels as of late 2026 and change often. Path Monthly cost level Time to first output PDP Law position Best fit Global subscriptions 20-30 USD per user Days Transfers need a legal basis; keep personal data out Internal work, English output, small teams Sahabat-AI managed (GPU Merdeka) Near global API parity, around IDR 7,500 per million tokens Weeks Data stays in Indonesia Customer-facing Bahasa and local-language service Self-hosted 70B node About 12,000 USD for an 8x H100 node, marginal tokens free 2-4 quarters including team Fully on-premises High-volume processors with MLOps staff Ref: costs compiled from vendor list prices and industry deployment guides cited in the Sources section. The work after the chat Whichever path a company picks, the subscription is the cheap part. The expensive part is the same for all three: connecting the model to the work. This is what the e27 column on Southeast Asian SMEs argued in November 2026. Most small businesses in the region already run on WhatsApp, the author noted, so the first AI purchase should not be a chatbot. It should be the system that takes what the chatbot hears and does the follow-up work: matching payments, updating the order sheet, filing the receipt. That integration work is where the real money goes, and it is also where vendors’ price lists go quiet. An integration project for a mid-sized retailer, connecting an AI service to its order flow and accounting, typically costs more in implementation labor in the first year than in model fees. The labor is billed in rupiah, by local firms, which is one more reason the domestic ecosystem around the models matters as much as the models themselves. Indonesia’s largest AI operator learned the same lesson at national scale. Indosat used AI-driven personalization on its own network and cut user churn by 50 percent over 90 days while lifting average revenue per user more than 6 percent, to 2.62 dollars. The capital expenditure saved through AI-optimized network planning reached about 86.5 million dollars. Asked what he would change, chief executive Vikram Sinha did not point at the technology. He said he would have pushed adoption across the organization harder and earlier. The machines were never the bottleneck. The habits were. Who should buy which path The choice reduces to three questions: where does the data live, what languages do the customers speak, and how many tokens per month. A company under 50 staff, doing internal work, with English-capable staff: buy the global subscriptions, keep personal data out, spend the savings on an integration project for the two workflows that actually cost time. A company with customer-facing service in Bahasa or local languages, handling personal data, with no MLOps team: use the managed sovereign route, and treat the language coverage of Javanese, Sundanese, Balinese or Batak as the feature that justifies the switch. A company processing millions of tokens daily, with or willing to build an operations team: self-host the open weights, accept the 12,000 dollar monthly floor, and stop paying per token at scale. The two adoption numbers will not converge on their own. The 54 percent learned to use AI because it was free and fun. The 14 percent will climb only when the work after the chat gets built, and the three price tags above are what the building actually costs in Indonesia in 2026. Data basis Adoption figures come from the e-Conomy SEA 2025 report by Google, Temasek and Bain & Company, released in November 2025, and from Microsoft adoption tracking cited in the accompanying research. Enterprise figures for Claude and Lazada come from the tool operators’ published usage data. Sahabat-AI capability, pricing and launch details come from the operators’ announcements of February 25, 2026 and industry deployment guides published in 2026. Cost estimates for the three paths use list prices and the industry-standard deployment benchmarks cited in the text; all are estimates and change with pricing updates. The Personal Data Protection Law framework follows UU PDP 27/2022, GR 33/2026 and OJK guidance as described in published legal analyses. The Fore Kopi breach report is from November 2026. Data as of November 2026. Limitations The 54 percent and 14 percent figures measure different populations with different survey methods, so the 3.65 ratio is indicative rather than exact. Token and subscription prices shift quarterly, and the self-hosting crossover point depends on utilization that no public dataset discloses. Production deployment counts for Sahabat-AI cannot be verified beyond public reviews. Regulatory interpretation of GR 33/2026 is still settling in practice. Sources BusinessWorld — e-Conomy SEA 2025 regional GenAI adoption figures, October 2026 VOI — Google Indonesia launch remarks, 18M daily images, November 2025 ITBeat — Sahabat-AI platform launch with Komdigi, February 2026 Technologue — GoTo and Indosat executive remarks on Sahabat-AI 70B, 2026 ITNews Asia — Indosat AI Grid results and Sinha remarks, 2026 Software Listing — SEA sovereign LLM deployment costs and benchmarks, 2026 e27 — SME chatbot workflows column, November 2026 CNBC Indonesia — Fore Kopi data breach report to Komdigi, November 2026 Lexology — Indonesia AI legal framework snapshot, 2026 About Global Infinity Labs Global Infinity Labs is an AI innovation company specializing in MCP data services. This article was prepared by our expert team using insights and data provided through our MCP platforms. Contact: support@globalinfinitylabs.com
More
Thailand’s EEC: How Manufacturers Secure Land Without a Nominee SeaPRwire

Thailand’s EEC: How Manufacturers Secure Land Without a Nominee

At the second Russian-Thai Investment Forum in Bangkok on 30-31 October, the Eastern Economic Corridor Office presented the corridor’s investment model to foreign capital, naming five sectors it wants filled: digital infrastructure and data centres, next-generation automotive, medical and life sciences, logistics and aviation, and the bio-circular-green economy. EECO executives also put the EEC Capital City project, known as EECiti, on the table as a joint-venture opportunity in residential, medical and wellness development. For manufacturers listening to that pitch, the immediate question is narrower than the five sectors: how a foreign-owned company lawfully controls industrial land in the corridor, and where the nominee shortcut that brokers still offer leads. The corridor itself has a fixed boundary. The EEC covers three eastern provinces, Chonburi, Rayong and Chachoengsao. Prachinburi sits outside it today; the EEC Policy Committee has approved its inclusion in principle, but expansion requires a royal decree, and in August 2026 the government confirmed publicly that the EEC still comprises three provinces. Sites marketed as “EEC-adjacent” in Prachinburi follow the general Thai rules, not the corridor’s special ones, and pricing a plot on assumed EEC status is the first avoidable mistake. Thai law starts from a prohibition. Section 86 of the Land Code B.E. 2497 bars foreigners from owning freehold land, and Section 113 targets the workaround: companies used to hold land for foreign beneficiaries. Two statutes open lawful doors around that bar. Under Section 27 of the Investment Promotion Act B.E. 2520, a company holding Board of Investment promotion can own land freehold in its own name for the promoted activity. Under Section 44 of the Industrial Estate Authority of Thailand Act B.E. 2522, operators inside an IEAT industrial estate gain the same statutory right. Cross-border counsel report structuring exactly this route for a precision-machinery investor in 2026, combining BOI Category 4.3 promotion with a 12-rai plot inside an IEAT free zone in the EEC, with title conveyed into the foreign subsidiary’s name within 21 days. The BOI route carries a maintenance obligation that gets less airtime than the right itself. If BOI promotion ceases, the company must dispose of the land within one year. That makes the durability of the promotion certificate a land-tenure issue, not just a tax issue, and it is why the due diligence sequence before wiring funds runs through verification of the Chanote title deed at the Land Office, confirmation of the plot’s colour-coded zoning under the master plan, and matching the manufacturing activity to an eligible BOI promotion classification. Funds for the purchase also need to enter Thailand in foreign currency with Foreign Exchange Transaction forms, so the paperwork trail exists before the deed does. Leasing is the second lawful route and the default for companies that cannot or do not want to hold title. A lease of land or buildings for more than three years must be in writing and registered at the Land Office to be enforceable for its full term; an unregistered long lease is a common and avoidable failure, because the tenant’s rights collapse to a short unregistered term exactly when a dispute or an exit arrives. EEC investment materials and forum reporting describe long-term lease arrangements, with terms reported at up to 50 years plus a 49-year extension for approved corridor investors, as part of the corridor’s package alongside tax incentives; investors should confirm the registered term and renewal mechanics plot by plot, since the registered document, not the brochure, is what a court reads. The lease-versus-buy decision also has lighter options that fit phased builds. Cross-border advisers working the corridor group the practical routes into five: buy a plot in an industrial estate, buy standalone industrial land, take a long-term registered lease, lease a ready-built factory, or commission a built-to-suit lease. A ready-built factory trades design control for speed to production, which suits a first-phase assembly line waiting on promotion approval; a built-to-suit lease moves design authority to the tenant but leaves delivery risk with the landlord, so the approval status of the building has to be negotiated explicitly. Acquiring an existing factory is the fastest route to an operating plant and the one where inherited permits deserve the most scrutiny, since the facility must be able to keep producing the buyer’s way lawfully after the deal. The third route is the corridor’s own promotion layer. EECO’s EECiti page lists the incentives the capital-city project offers, including the right to own land for business operations, corporate income tax exemptions or reductions, import duty relief, the right for foreign nationals to reside in Thailand for designated activities, and the right to use foreign currency. Strategic projects elsewhere in the corridor can carry corporate income tax exemptions of up to 15 years followed by a 50 percent reduction for five more, and qualifying foreign experts face a flat 17 percent personal income tax rate. The Thailand FastPass initiative launched in 2026 cuts approval times by up to 50 percent for qualifying investment projects. EECiti is also where the pipeline is concentrated. The project sits in Bang Lamung district, Chonburi, about 15 kilometres from U-Tapao airport, and covers roughly 14,619 rai, about 2,339 hectares, with development planned from 2026 to 2036 and an indicated investment value around 1.34 trillion baht. The design targets net-zero operation, with 100 percent clean energy, self-sufficient water management and green areas comprising no less than 30 percent of the total site. It runs on a public-private partnership split: the public sector handles land management and core infrastructure, and private capital takes residential, commercial and public-service projects. EECO has opened ten infrastructure investment areas to private participation, covering power and energy, water supply, wastewater and reuse, drainage and flood protection, solid waste, digital infrastructure and telecommunications, fire and disaster systems, transport, common utility conduits, and green areas. At the October forum, EECO’s EECiti division presented the infrastructure systems and the PPP framework directly to foreign investors. The nominee trap deserves its own paragraph because it is still being sold. Cross-border counsel describe a 2026 case in which local brokers proposed a 51/49 Thai limited company with nominee shares held by an offshore shell, a structure counsel assessed as fatally vulnerable under DBD Order No. 2/2569 and Land Code Section 113. Thai authorities actively check companies where foreigners hold more than 49 percent of shares if those companies buy land or villas. The penalty for using the structure is not a fine-and-continue arrangement; the land interest and the corporate structure are both exposed, and the fix is always a restructure into one of the lawful routes, which costs more than doing it properly the first time. Licensing around the land runs on separate clocks that investors should sequence, not discover. Factory operating approvals follow the Factory Act thresholds on machinery horsepower and workforce headcount, split across three licence classes. Environmental screening under ONEP regulations, at EIA, IEE or Code of Practice level depending on the project, has to be settled before plant architecture and equipment procurement are locked. Inside an IEAT estate the one-stop service centre can run building construction and factory operating approvals concurrently, which compresses the timeline; outside one, the approvals follow the standard provincial routes. The choice among the three routes follows from two facts: where the plant sits and whether the activity earns BOI promotion. A BOI-promoted factory in an IEAT estate holds the strongest title position, freehold in the company’s name under two statutes. A company without promotion, or on a site outside the estates, runs on a registered lease, which works well when the lease is registered and renewal and assignment rights are drafted for the building the tenant finances. An EECiti or special-promotion-zone project draws on the corridor layer, with land rights for business operations negotiated through EECO. What to watch next is the corridor’s own calendar: EECiti’s ten infrastructure packages move from the market-sounding stage toward PPP tenders through 2027, and each tender that lands will test how the land, licensing and incentive layers fit together for a foreign investor in practice. Sources EECO, “EECO Highlights Premier Investment Opportunities at the Russian-Thai Investment Forum 2026” (4 November 2026): five target sectors, EECiti joint-venture presentation and PPP framework. EECO, “EEC Capital City (EECiti)” project page: 14,619 rai site in Bang Lamung, Chonburi, 2026-2036 development period, around 1.34 trillion baht investment value, PPP model with ten infrastructure investment areas, and the incentive list including land ownership rights. Hey Thailand News, “Ford Expands in Thailand’s Eastern Economic Corridor”: EEC corporate income tax exemptions up to 15 years, 17 percent flat personal income tax for foreign experts, Thailand FastPass approval-time reductions, and the three-province corridor definition. Gentle Law IBL, “Thailand BOI Land Ownership & Factory License Guide 2026”: Land Code Section 86 prohibition, BOI Section 27 and IEAT Section 44 freehold rights, the one-year disposal rule, DBD Order No. 2/2569 nominee case, and the due diligence sequence. Lex Bangkok, “Finding Factory Land or a Factory to Lease in the EEC”: three-province boundary and Prachinburi status as of August 2026, lease registration above three years at the Land Office, and the buy/lease route comparison. Disclaimer: This article is provided for general information only and does not constitute legal advice. About Global Infinity Labs: Global Infinity Labs is an independent research desk covering cross-border commerce, logistics and regulation. The desk publishes field notes for operators, compliance teams and investors working across Asian and Gulf markets. Email: support@globalinfinitylabs.com Submission portal: https://newsroom.seaprwire.com
More
UAE Stablecoin Register Adds a New Layer to GCC Payment Licensing SeaPRwire

UAE Stablecoin Register Adds a New Layer to GCC Payment Licensing

On 9 November 2026, Liv, the digital bank of Emirates NBD, put USDU, the first US dollar-backed stablecoin registered by the Central Bank of the UAE, into its customer app. Universal Digital Intl Ltd issues the token under the central bank’s Payment Token Services Regulation, with one-to-one dollar reserves held in UAE banks and verified through monthly public attestations. Aquanow ME FZE, a broker-dealer licensed by Dubai’s Virtual Assets Regulatory Authority, supplies the trading infrastructure, and the bank is running a 90-day zero-fee campaign to pull customers onto the new rails. The launch is a product event, but the layer underneath it, a central-bank register of payment tokens with named issuers, is what changes the licensing map for payment firms looking at the Gulf. That register has explicit rules. Under the Payment Token Services Regulation, a merchant in the UAE may accept a virtual asset as payment only if it is a Dirham Payment Token issued by a licensed issuer, or a Foreign Payment Token from a registered issuer used to purchase a virtual asset or virtual asset derivative. As of September 2026, one foreign payment token held CBUAE registration: USDU itself, issued by the ADGM-licensed Universal Digital. On the dirham side, four tokens were licensed or approved: AE Coin, Zand AED, DDSC, and RAKBANK AED on an in-principle basis. Registration is therefore not a formality attached to any stablecoin; it is a closed list, and market access for payment use runs through it. The distribution choice is the other half of the UAE model. USDU reached customers inside a banking app rather than through a standalone exchange, with the issuer, a VARA-licensed infrastructure firm and one of the region’s largest banks each holding a distinct regulated role. The UAE splits financial-authorisation work across several regulators: the CBUAE for payments and payment tokens at federal level, the Capital Markets Authority for investments outside the financial free zones, VARA in Dubai, and the DFSA and FSRA inside DIFC and ADGM. The federal baseline matters here: Federal Decree-Law No. 6 of 2025 sets out, in its Article 61, the financial activities that require a central bank licence, a list that includes payment services using virtual assets, stored value facilities, and retail payment and digital money services. A free-zone trade licence authorises none of this on its own; the token’s function decides the regulator, with payment use falling to the central bank. Foreign payment firms mapping the UAE therefore need to answer one question before any other: which register does the planned activity sit on. Saudi Arabia’s permit tiers are easiest to read through the companies already licensed. SAMA’s register lists established operators alongside banks: Geidea, for example, appears as a Payment Institution with a licence running to November 2030, a firm that reported being the only non-bank acquirer licensed in the Kingdom in March 2021 and added a payment gateway licence in 2023 and an e-commerce licence in 2024, and which announced certification in September 2026 to route e-commerce payments to the National Payments Gateway alongside MPoC certification for its SoftPOS application. The step pattern, acquiring first, gateway second, e-commerce third, shows how the Kingdom’s authorisations are stacked per activity rather than granted as a single blanket payment permission. Saudi Arabia runs the opposite structure: one statute, one regulator. The Law of Payments and Payment Services, enacted by Royal Decree M/26 and implemented through regulations effective 13 June 2023, gives the Saudi Central Bank sole licensing and supervisory authority over payment systems and payment service providers across the Kingdom. The entry gate is binary in design: a bank-PSP route under banking law, or a non-bank authorisation route for payment institutions and e-money institutions under the payments law. SAMA has been layering specialist permits on top, and the tiering shows up in company disclosures: SurePay, a payments subsidiary, holds a Small Payment Institution licence from SAMA, a point-of-sale payment service provider permit, a SoftPOS variant of the same permit, and, as disclosed in November, the Payment Technical Service Provider technical certificate for e-commerce payments. A foreign operator entering Saudi Arabia pairs the SAMA authorisation with a MISA investment licence, and the open banking framework transitioned to a licensing regime in March 2026, adding account information and payment initiation services to the regulated set. Qatar’s layer moved in a different direction in September. On 1 September 2026, the Qatar Central Bank enabled licensed payment service providers to open accounts at the central bank and settle directly through the Qatar Real-Time Gross Settlement system, without routing through intermediary banks. SADAD Payment Solutions and Dibsy were the first two PSPs to join. The central bank framed the move as cost reduction for providers and merchants, a shorter payment cycle with fewer parties, and stronger direct oversight of settlement. Two weeks later, on 15 September, Qatar and Saudi Arabia switched on reciprocal acceptance of Qatar’s Himyan cards and Saudi Arabia’s mada cards across both markets’ merchant and ATM networks. Non-bank payment firms in Qatar now sit closer to the settlement core than their counterparts did a year ago, with the operational obligations that proximity brings. Entry testing in Qatar runs through the central bank’s fintech routes as well. The Regulatory Sandbox and the Express Sandbox, a fast-track launched in May 2024 with a reduced testing period and streamlined evaluation for solutions that demonstrate product readiness, are the supervised paths toward full authorisation, and the country’s instant payment rail Fawran and the Qatar Mobile Payment System QR standard give licensed firms settlement-grade infrastructure to build on. A firm that graduates from sandbox testing into the licensed perimeter is then a candidate for the direct RTGS participation described above, which is the sequence QCB appears to be encouraging. Set side by side, the three markets ask different things of a foreign payment firm. Saudi Arabia asks for one authorisation from one regulator, with permit tiers matched to activity scope, and the paperwork concentrated at entry. The UAE asks for correct register placement first, since the same token can be a payment question, a securities question or a virtual-asset question depending on use, and distribution increasingly runs through bank partnerships with separately licensed infrastructure providers. Qatar asks for capital and operational readiness befitting direct settlement participation, plus merchant-level approvals that add onboarding friction inside the licensed perimeter. None of the three publishes a licence shelf-price; the real gate is which activities the regulator maps onto which category. Two developments deserve a line in any entry plan. First, registers are moving targets in both directions: the UAE’s payment token list grew twice this year and each addition, licensed or registered, redefines who may transact and in which direction. Second, oversight is climbing toward the infrastructure layer across the region, a pattern consistent with the World Payments Monitor’s report that SAMA has opened a consultation on a modernised oversight framework for payment systems and their operators, which would introduce a formal systemically important payment system classification and an annual self-assessment duty; the report gives a closing date of 27 November 2026, and SAMA’s own consultation page had not been retrievable at the time of writing, so that item should be confirmed against SAMA’s announcements before being relied on. What to watch through the next quarter is straightforward. The CBUAE register will show whether a second foreign payment token and the four in-principle dirham tokens convert to full licences. QCB has said additional payment service providers can join QA-RTGS in accordance with applicable requirements, so the first-two list will grow. And any SAMA publication on payment system oversight, whether a consultation document or a framework, will define which Saudi operators sit in the systemically important tier, and with it the self-assessment calendar that comes attached. Sources Emirates NBD (Liv), press release of 9 November 2026: USDU as the first CBUAE-registered USD-backed stablecoin under the Payment Token Services Regulation, issued by Universal Digital with 1:1 reserves and monthly attestations, distributed through Liv with Aquanow’s VARA-licensed infrastructure. ARP Digital, “Stablecoin AML: What Compliance Teams Actually Have to Do”: Payment Token Services Regulation merchant-acceptance rule, the September 2026 register position with one registered foreign token and four dirham tokens, and source citations to the CBUAE regulation. World Payments Monitor, “Saudi Arabia Payments & Fintech Regulation” jurisdiction brief: Law of Payments and Payment Services (Royal Decree M/26) and implementing regulation effective 13 June 2023, the bank-PSP and non-bank PI/EMI routes, the March 2026 open banking licensing transition, and the SAMA oversight framework consultation reported with a 27 November 2026 closing date. Gulf Times, “Qatar Central Bank enables direct QA-RTGS access for payment service providers”: direct settlement access from 1 September 2026 with SADAD Payment Solutions and Dibsy as the first two PSPs, and QCB’s stated cost, simplification and oversight rationale. The Saudi Standard, “Sure Global Tech Subsidiary Wins E-Commerce Payment Cert”: SurePay’s SAMA Small Payment Institution licence, POS and SoftPOS permits, and the November PTSP e-commerce certificate as a worked example of Saudi permit tiering. Payment Review, “Geidea Review” (November 2026): SAMA register entry for Geidea as a Payment Institution with licence expiry 4 November 2030, the 2021 acquiring licence, 2023 gateway licence, 2024 e-commerce licence, and September 2026 National Payments Gateway certification. Uniwide Advisors, “Crypto Licence in the UAE”: Federal Decree-Law No. 6 of 2025 Article 61 list of CBUAE-licensed activities, including payment services using virtual assets and stored value facilities, and the four-regulator split. Disclaimer: This article is provided for general information only and does not constitute legal advice. About Global Infinity Labs: Global Infinity Labs is an independent research desk covering cross-border commerce, logistics and regulation. The desk publishes field notes for operators, compliance teams and investors working across Asian and Gulf markets. Email: support@globalinfinitylabs.com Submission portal: https://newsroom.seaprwire.com
More
EU AI Act After the Omnibus: The Two Deadlines That Now Matter SeaPRwire

EU AI Act After the Omnibus: The Two Deadlines That Now Matter

The European Commission has fixed two dates for the EU AI Act’s high-risk regime that replace every calendar a compliance team drew up before July. Standalone high-risk systems listed in Annex III must meet the Act’s full obligations from 2 December 2027. High-risk AI embedded as a safety component in regulated products, the Annex I list, follows from 2 August 2028. Both dates come from Regulation (EU) 2026/1744, the Digital Omnibus amendment, which entered into force on 27 July 2026. The old date, 2 August 2026, has already passed, and it passed without most high-risk obligations applying. The amendment did not appear without warning. The Commission proposed the Digital Omnibus on AI on 19 November 2025, when two gaps had become obvious. The harmonised standards that providers need in order to demonstrate conformity with the high-risk requirements were not finalised; conformity assessment against standards that do not exist is a paper exercise. And several member states had not yet stood up their national supervisory authorities, so enforcement on the original schedule would have had nowhere to land. The European Parliament adopted the amendment on 16 June 2026 and the Council followed on 29 June. What applies today is the part many organisations misread. The prohibitions on practices such as social scoring, untargeted facial scraping and emotion recognition in workplaces have applied since 2 February 2025. The general-purpose AI model obligations, covering technical documentation, downstream information and a copyright policy, have applied since 2 August 2025. The Article 50 transparency duties, which require people to be told when they are interacting with an AI system and require synthetic content to carry machine-readable marking, have applied since 2 August 2026. The Commission’s own AI Act overview page lists these as current, and none of them moved. The Article 50 layer is worth spelling out for B2B vendors, because it touches products that never come near the high-risk list. A chatbot interface must disclose that the user is talking to a machine. Emotion recognition and biometric categorisation systems must notify the people exposed to them. Generative systems must mark synthetic output in machine-readable form, deepfakes must be labelled visibly, and text published to inform the public on matters of public interest must be disclosed as machine-generated. Since 27 July 2026, consumer representative actions under the Injunctions Directive also cover the AI Act, so an incomplete disclosure regime is no longer only a supervisory-authority matter; qualified entities can pursue it in court. GPAI duties also carry a systemic-risk tier. Models whose training compute reaches the threshold set in the Act are presumed to carry systemic risk, which brings security evaluations, risk mitigation, incident reporting and cybersecurity obligations on top of the baseline transparency and copyright duties, under Article 51 of Regulation (EU) 2024/1689. The Commission’s AI Office supervises this tier directly. December 2026 adds two entries of its own. From 2 December 2026, the Act bans AI that generates non-consensual intimate imagery of real people, the category that covers so-called nudification apps, and AI that generates child sexual abuse material. On the same date, providers of generative systems that were already on the market before 2 August 2026 must complete their catch-up on Article 50(2): output must carry machine-readable marking and be detectable as artificially generated or manipulated. Systems placed on the market after 2 August 2026 never had a transition; the marking duty bound them at launch. The December 2027 date is where the heavy documentation lives. Providers of Annex III systems, the standalone category that includes recruitment and CV-screening tools, credit scoring, biometric categorisation, critical infrastructure management, education assessment and law enforcement uses, must complete a conformity assessment under Article 43, operate a quality management system under Article 17, compile technical documentation against Annex IV, register the system in the EU database under Article 71, affix CE marking and appoint an EU authorised representative if they sit outside the Union. None of these requirements was softened. The omnibus moved the finish line and left the course unchanged. Deployers carry their own list on the same date. Human oversight measures under Article 26, logging of inputs and outputs, informing employees where AI-assisted decisions affect them, and, for public body deployers, a fundamental rights impact assessment. A vendor that treats December 2027 as a provider-side problem will find its customers asking contractual questions well before the regulator does, because deployer documentation obligations depend on information only the provider can supply. The Annex I track runs on a longer fuse for a structural reason. Embedded high-risk AI, the kind that forms a safety component of machinery, medical devices, vehicles or toys, is assessed through the conformity procedures those product regimes already run. Two regulatory frameworks have to line up in a single assessment, which is why the legislature added 24 months rather than 16. Machinery is now routed through the Machinery Regulation, and the Commission must request harmonised standards that cover the AI Act and sector law together rather than in parallel documents. Two edge cases are worth checking against inventory before any other work. Systems already on the market before their new compliance date are only caught if they are substantially modified afterwards. And AI used by public authorities has until 2 August 2030, the longest runway in the Act. Beyond dates, the omnibus recalibrated several duties. Article 4 still requires providers and deployers to take measures supporting AI literacy among staff, but it no longer guarantees a specific level of competence for every individual, and the Commission and member states now carry a duty to support organisations, smaller ones in particular, with practical guidance. Small mid-cap enterprises, defined as companies with fewer than 750 employees and either up to 150 million euros in turnover or a balance sheet up to 129 million euros, gain access to the simplified documentation regime, sandbox priority and lower fine caps that SMEs already had. The AI Office gained exclusive supervision of AI systems built on a general-purpose model from the same provider or corporate group, and of AI integrated into very large online platforms and search engines. The amendment also postponed the national regulatory sandboxes chapter alongside the two high-risk blocks, so member-state sandbox programmes built against the old dates need their own revised schedules. The penalty architecture did not move. Violations of the prohibited practices can reach 35 million euros or 7 percent of global annual turnover, whichever is higher, and general-purpose AI breaches can reach 15 million euros or 3 percent, under Article 99 of Regulation (EU) 2024/1689. Member state authorities enforce most provider and deployer duties; the Commission enforces the GPAI chapter. For B2B software providers selling into the EU, the practical sequence is short. First, classify: most B2B tools are minimal-risk or fall under Article 50 transparency, and only a minority sit in Annex III. Second, date-stamp each system against the correct line, because a recruitment-screening module and a chatbot in the same product suite now live under different calendars. Third, treat the December 2026 marking catch-up as the first hard checkpoint; it is six weeks away and applies to systems already shipped. Fourth, watch notified body capacity, since the December 2027 concentration of conformity assessments is the kind of bottleneck that forms quietly and prices itself into audit slots. What to watch through 2027 is narrower than the topic suggests. The harmonised standards programme at CEN and CENELEC, the Commission’s implementing acts, and member states’ designation of supervisory authorities are the three inputs that determine whether December 2027 holds. The omnibus was written because those inputs slipped once. The dates are now fixed in law, which cuts both ways: no further slippage is scheduled, and none is assumed. Sources European Commission, “AI Act” policy page: confirms high-risk obligations from 2 December 2027, transparency rules from August 2026, and the December 2026 prohibition added by the AI Omnibus. I-Scoop, “The EU AI Act after the Digital Omnibus”: timeline as of November 2026, including Regulation (EU) 2026/1744 publication on 24 July 2026 and entry into force on 27 July 2026, the 2 August 2030 public-authority date, and small mid-cap thresholds. Regulation AI, “EU AI Act Omnibus 2026: New Compliance Deadlines”: provider and deployer obligation lists for the 2 December 2027 and 2 August 2028 dates, with Articles 17, 43, 48, 71 and 26 references. AI Act Insight, “EU AI Act: the complete guide”: consolidated timeline showing Article 50 transparency applied from 2 August 2026 and the 2 December 2026 marking catch-up under Article 50(2). Regulation (EU) 2024/1689 (EU AI Act), EUR-Lex consolidated instrument: penalty tiers under Article 99. Disclaimer: This article is provided for general information only and does not constitute legal advice. About Global Infinity Labs: Global Infinity Labs is an independent research desk covering cross-border commerce, logistics and regulation. The desk publishes field notes for operators, compliance teams and investors working across Asian and Gulf markets. Email: support@globalinfinitylabs.com Submission portal: https://newsroom.seaprwire.com
More
Building Costs Re-Accelerate in Australia as NZ Consents Sink SeaPRwire

Building Costs Re-Accelerate in Australia as NZ Consents Sink

Construction costs across Australia rose 1.0 percent in the September quarter, according to the latest Cordell Construction Cost Index published by Cotality on 21 October. The quarterly figure is five times the 0.2 percent recorded at the start of 2026, and annual growth lifted to 3.2 percent from 2.8 percent the quarter before. Both readings say the same thing: the cost slide that opened the year is over, and escalation is back inside the range that builders priced against in late 2024 and 2025. John Bennett, Cordell Costings Estimation Manager at Cotality, called the soft opening months an anomaly and said the return to 1.0 percent quarterly growth brings cost escalation back in line with the levels recorded over the past two years. His firm’s state table shows the recovery is broad but not uniform. Western Australia led with a 1.1 percent quarterly rise, Victoria and South Australia posted the lowest growth at 0.9 percent, and New South Wales and Queensland each matched the national 1.0 percent. Cotality’s longer view softens the headline: every state remains below its five-year average growth rate, so this is re-acceleration from an unusually low base, not a return to the post-pandemic spiral. The official statistics line up with the index. The Australian Bureau of Statistics released its Producer Price Indexes for the September quarter on 30 October, reporting that output prices for building construction rose 1.3 percent in the quarter, driven by non-residential construction, with residential building construction output up 1.1 percent. Final demand prices rose 1.1 percent for the quarter and 3.6 percent through the year. The ABS attributes part of the residential rise to skilled labour shortages feeding wages, plus elevated freight and transport costs on imported materials. The input side tells a sharper story. Master Builders Australia flagged the ABS data on 30 October, noting that input costs for house construction rose 3.9 percent over the year to September, the fastest annual pace since the September quarter of 2023. The categories doing the damage are narrow but heavy: electrical equipment up 14.5 percent over the year on high global copper prices, plumbing products up 9.4 percent, other metal products including copper pipe up 5.0 percent, and ceramics up 4.1 percent. Output prices charged for a finished house rose 5.7 percent nationally, 7.8 percent in Queensland and Western Australia, and 10.7 percent in Tasmania. Trade-level movements in the Cotality index were mixed. Plastics saw broad-based increases in the quarter and plasterboard recorded smaller uplifts, and Bennett noted that some costs associated with the Middle East conflict may now be flowing through to material prices. Suppliers appear to be returning to standard pricing cycles after an unsettled stretch, which matters for builders already running thin margins on overheads. Two technical points in the ABS release affect how those percentages should be read. Fuel excise and other government taxes sit outside the Producer Price Indexes, so the unwinding of the federal fuel excise relief in July and August did not push the PPI construction series up directly, though elevated transport and freight charges on imported materials did register, and oil prices eased over the September quarter itself. The ABS has also postponed the 2026 weight update and will run a staged reweight by industry, starting with a full reweight of the Construction index at the December quarter release. Index users who write escalation clauses against these series should expect a basket change, not just new numbers, in the January 2027 print. The consumer-side reading is softer than the input side, which is where negotiation room shows up. The ABS Consumer Price Index has new dwelling prices up 4.7 percent over the year to September, down from 5.4 percent in the year to August, and flat for the month: some project builders raised base prices while others widened discounts and promotions as new demand slowed. A buyer comparing quotes in the current quarter is negotiating against builders whose cost base is rising faster than the prices they can charge. Across the Tasman, the same cost pressure is colliding with a volume contraction. Stats NZ data reported by The Post on 3 November shows the total value of non-residential building work consented over the year ending September fell 3.1 percent to 8.71 billion New Zealand dollars, with factories and industrial buildings followed by offices recording the biggest declines. Non-residential construction prices were up 2.2 percent over the year to June 2026, which means the value decline is understated in volume terms: after adjusting for building cost inflation, annual consent volumes slipped below their May 2026 and May 2020 results to a 12-year low, according to Infometrics chief forecaster Gareth Kiernan. September activity alone was down 18 percent on the same month a year earlier. The residential picture in New Zealand runs the other way, with a catch attached. Consents for 41,091 new homes were issued over the year ending September, an 18 percent increase on the prior year. But September’s new-home consent count fell 7.9 percent, the fourth decline in five months, and Westpac senior economist Satish Ranchhod read the series as a peak with planned work easing off. Higher interest rates, a soft housing market and picked-up building costs are all cited as brakes on the next wave of projects. Inside the weak non-residential series there are pockets of strength that matter for fit-out and services contractors planning capacity. Education, hospital and storage building consented ahead of expectations in the year to September, and major consents for the Christchurch Men’s Prison and Dunedin Hospital projects are expected to lift activity over the coming six months. The public-side numbers are a different story: year-end growth in public sector consent value weakened through 2026 to a 17-month low in September, and private sector consent value slipped back into negative territory after a positive August. Kiernan also points to persistently high fuel prices holding up construction costs wherever heavy machinery is involved, with the broader dampening effect on the recovery limiting demand for commercial and industrial space. Perth shows how the two forces can squeeze a project from both ends at once. Cotality’s Home Value Index for November 2026 has Perth dwelling values down 5.7 percent over the quarter, with the median unit at about 692,000 Australian dollars, while Western Australia recorded the strongest construction cost rise of any state at 1.1 percent for September. Developers interviewed for the Australian Housing and Urban Research Institute’s latest delivery report said medium and high-density projects in the city largely stack up only at the top of the market. A feasibility model built on falling sale prices and rising build costs produces a widening gap, and Precinct Property’s decision to defer its Downtown Carpark redevelopment in Auckland, citing higher interest rates, construction cost risks and weak conditions, is the same arithmetic in a different city. For contractors and developers budgeting the next tender cycle, the quarter suggests three concrete adjustments. Fix contract prices early where quotes remain valid, because the 0.2 percent quarters are behind rather than ahead. Read rise-and-fall clauses with the 3.9 percent input trend in mind, particularly on copper-dependent electrical packages and plastics, and confirm whether supplier fuel surcharges still embedded in quotes will come off if those charges end. And hold contingency as a budget line rather than a discount, since the state results show cost paths diverging by up to 0.2 points between the strongest and weakest markets in a single quarter. What to watch next is defined on the calendar. The ABS has scheduled a reweight of the Construction index for the December quarter release on 29 January 2027, following the postponed 2026 weight update, which will reset the basket that the headline percentages are calculated on. Cotality’s December quarter index will show whether the fourth quarter repeats September’s pace or the year-end softens again. Between those two prints, tender margins, not headline inflation, remain the number that decides which projects proceed. Sources InfrastructureNews, “Australia construction costs rise 1.0% as momentum returns” (2 November 2026): Cotality Cordell Construction Cost Index September quarter figures, state breakdown, and John Bennett commentary. Australian Bureau of Statistics, “Producer Price Indexes, Australia, September 2026” (released 30 October 2026): output of building construction +1.3% for the quarter, final demand +1.1%/+3.6%, and the December quarter construction reweight schedule. The Post, “Non-residential building consents fall to 12-year low” (3 November 2026): Stats NZ consent value of NZ$8.71 billion, the 2.2% construction price series, and Infometrics and Westpac commentary. Homebuilds, “Building Costs Rise at Fastest Pace in Three Years”: ABS input cost detail by category and Master Builders Australia’s 30 October flag. The West Australian, “New apartments will not get cheaper until building them does”: Western Australia quarterly rise, Cotality Home Value Index for November 2026, and AHURI delivery report context. Disclaimer: This article is provided for general information only and does not constitute legal advice. About Global Infinity Labs: Global Infinity Labs is an independent research desk covering cross-border commerce, logistics and regulation. The desk publishes field notes for operators, compliance teams and investors working across Asian and Gulf markets. Email: support@globalinfinitylabs.com Submission portal: https://newsroom.seaprwire.com
More
Shaping the Future Together: Voices from Around the World at the III International Symposium ‘Creating the Future’ in Russia SeaPRwire

Shaping the Future Together: Voices from Around the World at the III International Symposium ‘Creating the Future’ in Russia

Moscow, Russia – November 15, 2026 – (SeaPRwire) – The III International Symposium ‘Creating the Future’ brought together more than 8,000 participants from 109 countries at the Russia National Center in Moscow. It is a key event within the global ecosystem of the Open Dialogue “The Future of the World: A New Platform for Global Growth”. Symposium ‘Creating the Future’, photo: press service of the Russia National Center Public figures, scientists, futurists and representatives of culture, media and business exchanged ideas and experience to develop positive scenarios for the future. The President of Russia sent a message of greeting to participants, noting the event’s international standing and the practical value of bold, unconventional ideas and solutions. Russian economist Maxim Oreshkin formally opened the Symposium on 12 November. Oreshkin said representatives of 19 more countries had joined the project in 2026. He announced the opening of submissions for the third season of the Open Dialogue, which invites experts, businesspeople and researchers worldwide to submit essays on humanity’s future. In 2026, the competition received more than 1,600 entries in 20 languages from over 100 countries. Submissions for the third season are accepted at www.dialog.russia.ru/en/. Inviting authors from different professions and cultures to take part, Oreshkin said the new cycle would focus on people in a changing technological world. Its four themes are “Human Development”, “Health Management”, “Life Around Us” and “Content in a New Era”. The most compelling ideas will inform a new report, and the authors of the best submissions will be invited to Moscow in April to discuss the results. The programme comprised more than 40 lectures, discussions, meetings and training sessions across three tracks: “Human and Values”, “Progress”, and “Culture and Art”. Politicians, scientists, futurists and representatives of creative industries, media, civic and youth organisations took part in interdisciplinary exchanges. Natalia Virtuozova, Director General of the Russia National Center, described it as a space for national pride and equal international dialogue. She welcomed the participation of 109 countries and said: “Progress represents the direction in which we are moving: we are building up our technological potential, while always remembering that every achievement serves people.” She stressed that culture, art and traditional values provide the foundation for the future, and hoped each participant would take away new ideas and discoveries. The II Museum Congress, “The Museum – A Place for Shaping the Future”, was among the Symposium’s principal events. On 12 November, representatives of 75 Russian and eight overseas museums discussed the transformation of museums in the digital age, connections between young people and history, national traditions and contemporary challenges. Moaza Matar, Director of Curatorial Affairs and Collection Management at the Zayed National Museum in the UAE, said that meeting representatives of other museums had helped her team exchange practices and consider how to make knowledge accessible, especially to students and younger generations. She saw opportunities to continue cooperation with the Russia National Center. For the first time, the Symposium hosted the Gaming Lobby “Gaming Session of the Future”, bringing together game industry experts, developers, gamers and esports fans. Participants discussed video games as a cultural phenomenon and explored Russian releases, including DISTORTION, a story-driven action RPG presented at the event. The developers shared initial testing results, analysis and plans for the game’s universe. The open discussion “Drafts of the Future” brought science-fiction writers and futurists together. The programme culminated in the II All-Russian Science Fiction Award, established by the Russia National Center and the Union of Writers of Russia to renew wider interest in the genre as a stimulus for scientific and technological progress. More than 120 works published in 2025 were nominated. Russian and international writers, translators and artists won awards in five main and two special categories. The new “International Recognition” category aims to support the genre worldwide and introduce Russian authors to overseas audiences. One of its recipients, science-fiction writer and President of the China Science Fiction Writers Association Chen Qiufan, said that examining different periods helps people understand possible futures. He described discussions among writers, futurists and scientists as important for future discoveries, and highlighted science fiction’s role in developing young people’s imagination and creative thinking, alongside support for authors in Russia and China. The Symposium also addressed the balance between innovation and tradition. The discussion-performance “The Power of Craft: Preserving Heritage for the Future” considered folk crafts as a link between generations and part of the cultural heritage. A new project for young people, the Young Artisan School, was announced for 2027 at the Russia National Center. It will introduce children to traditional crafts and making objects by hand. During the Symposium, the Center presented a premiere run of “Khorovod. The Energy of Unity”, a musical and choreographic parable featuring more than 600 performers from regions across Russia. The production introduced participants to the diverse cultural traditions of the country’s peoples and became one of Russia’s major cultural events of 2026. Social Links Telegram: https://t.me/gowithRussia Media Contacts Brand: Russia National Centre Contact: Olivia Stone Email: pressa@russia.ru Website: https://en.russia.ru
More
Dubai Moves Three Trade Shows to Expo City for 2027 SeaPRwire

Dubai Moves Three Trade Shows to Expo City for 2027

On 3 November, Gulfood Manufacturing opened across the eight numbered halls at the Dubai World Trade Centre, together with Za’abeel Halls 2, 3, 4 and 6, the three Sheikh Saeed Halls, Sheikh Maktoum Hall, Sheikh Rashid Hall, the Trade Centre Arena and the Al Multaqua Ballroom. That is most of the downtown campus. When the same exhibition returns on 1 November 2027, it will run in the North and South Halls at Dubai Exhibition Centre, at Expo City in Dubai South. Two campuses, one operator Dubai Exhibition Centre is owned and operated by Dubai World Trade Centre, the same organisation that runs the downtown complex on Sheikh Zayed Road. They are not competitors. The operator describes the pair as giving organisers “the flexibility to accommodate different event formats and requirements,” adding that together they allow several major exhibitions to run concurrently. Exhibitors are therefore not choosing between rival landlords. They are being placed. And the 2027 placement is already published. Three shows change campus Set the organiser’s calendar for the rest of 2026 against the 2027 schedule published for Dubai Exhibition Centre, and at least three exhibitions cross over. Beautyworld Dubai moves from 6-8 October at the downtown site to 19-21 October at Expo City. Gulfood Manufacturing moves from 3-5 November to 1-3 November, keeping its month and losing its address. Middle East Energy moves from 1-3 September to 11-13 May, changing the month and the campus at once. That last one carries the most consequence. A September Dubai trip and a May Dubai trip are not the same commercial trip, and a stand plan built around the September cluster will not transfer. The three join a group that has already crossed. GITEX Global runs 8-11 December 2026 at Dubai Exhibition Centre, after a one-day summit at the downtown venue on 7 December. Automechanika Dubai took the Expo City halls on 10-12 November. Gulfood and World Health Expo were the two shows the operator used to demonstrate that the expanded venue could carry real scale: Gulfood 2026 spread across both campuses and more than 280,000 square metres, while World Health Expo recorded more than 235,000 professional visits and 4,300 exhibitors from over 180 countries, its largest edition in five decades. What the new halls actually are Phase 1 of the expansion was delivered in early 2026, built in 14 months. It took purpose-built indoor event space to 140,000 square metres with a daily visitor capacity of 50,000. That is the floorplate GITEX will use in December. Phase 2 is the phase that governs 2027 scheduling. Works moved fully into it in October 2026, covering extended concourses, new East Halls and multi-storey car parks. Completion takes indoor event space to 160,000 square metres, daily capacity to roughly 60,000, and adds around 7,000 parking spaces. A new eastern access road forms part of the same phase, described by the operator as a critical enabling project that keeps access open to the venue, the Expo City hotels and the basement facilities while construction continues. The end state, scheduled for 2031, is 180,000 square metres across 1.2 kilometres from north to south, a daily capacity of 65,000, and room for more than 20 events at once. The programme sits inside an AED 10 billion masterplan, with the operator targeting an annual economic contribution of AED 54 billion by 2033. The operator’s own Q4 calendar already lists more than 100 exhibitions, conferences and business events across the two venues between September and December 2026. Where the split falls The split follows floorplate size. The downtown site carries high-frequency, mid-sized trade shows and the parallel events that bunch into a single week. Between 3 and 5 November it hosted four at once, Gulf Host, Gulfood Manufacturing, ISM Middle East and Private Label Middle East, spread across the numbered halls, Za’abeel, Sheikh Saeed and the Arena. Expo City takes the shows that need contiguous floor space at a scale the downtown site cannot supply. GITEX, Gulfood, World Health Expo, Automechanika, and now Beautyworld, Middle East Energy and Gulfood Manufacturing. The sorting rule is capacity, not industry. Four lines to redo in the budget The campus change is not cosmetic. It moves four cost lines, and each one behaves differently. Transit is the smallest change and the easiest to get wrong. Both sites sit on the Red Line. The downtown campus has its own station, World Trade Centre, with the concourse feeding directly into the halls. Dubai Exhibition Centre is served by Expo 2020 station, a five to ten minute walk from the entrances, with trains every three to four minutes. During GITEX the organiser runs direct shuttles between the two venues. Dubai International Airport to the downtown site is about 20 minutes by metro, per the event’s own travel guidance. Hotels are the line that changes most. The downtown campus has properties on the plot, including Novotel and Ibis hotels, which is why a stand crew can walk from room to hall in minutes. Expo City’s expansion plan includes a 300-key hotel, but that is a planned facility rather than an operating one. Until it opens, teams working the Expo City halls are choosing between accommodation at Expo City and a daily commute from the city. Parking changes shape, not size. Dubai Exhibition Centre currently offers more than 2,300 visitor bays with shuttle connections to the Arrivals Plaza, plus park-and-ride at four Red Line stations. Phase 2 adds the multi-storey capacity, around 7,000 spaces, which matters most during build-up and breakdown when vans and crates arrive in waves. Arrival path is the long-horizon item. Dubai Exhibition Centre sits beside Al Maktoum International Airport, being developed as one of the world’s largest passenger hubs. That will not settle a 2027 budget, but the visitor flying in for Expo City will not land where the downtown visitor lands. The construction caveat Phase 2 is being built while exhibitions run. The operator states that Dubai Exhibition Centre remains fully operational throughout, and the 2026 calendar is consistent with that, with Gulfood, World Health Expo, Automechanika and others running through the works. Access routes, gate assignments and drop-off points can change between editions of the same show. For a freight forwarder or a stand contractor, last year’s site plan is not a reliable document. Ask for the current one at the point of booking, then ask again around 60 days before build-up, when the site configuration for that specific show is fixed. What to settle before signing Four questions decide whether a 2027 stand budget survives contact with the calendar. Which campus. Read the 2027 schedule, not the 2026 one. Three exhibitions have already moved and the published 2027 list is the only reliable source. Which month. Middle East Energy moved from September to May, so any co-location plan needs rebuilding from the dates up. What the hotel block assumes. Downtown and Dubai South are different hotel markets at different rates and different journey times. A block priced against the downtown campus will not hold at Expo City. What the freight route looks like this year. Phase 2 works, the eastern access road and the new car parks all touch vehicle movement. Confirm the route when the freight is booked, then reconfirm close to the show. One direction The expansion is funded, phased and published, and the exhibitions that moved have not moved back. Two of the three shows changing campus in 2027 keep their month, which suggests the organiser is treating the shift as a capacity transfer rather than a calendar reset. For exhibitors, the useful response is not to wait for the final configuration that arrives in 2031. It is to treat 2027 as a year in which both campuses are live, price the campus variable into stand and travel budgets now, and check the 2027 calendar instead of the one that ran this year. Sources ## Sources Dubai World Trade Centre: Q4 business events outlook as DEC expansion advances, 9 September 2026. Dubai World Trade Centre: Events calendar across two venues through year-end 2026, 19 August 2026. Dubai Exhibition Centre: Expansion plans and phase schedule. Dubai Exhibition Centre: Venue guide and transport access. Dubai Exhibition Centre: What’s on, 2026-2027 calendar. Dubai Exhibition Centre: About us, ownership and operator. Dubai World Trade Centre: What’s on, hall allocation by event. GITEX Global 2026: Plan your visit, venues and transport. Dubai Exhibition Centre: Gulfood 2026 across two venues, 2 February 2026. Dubai Exhibition Centre: WHX 2026 largest edition in its history, 13 February 2026. Dubai Exhibition Centre: Phase 2 expansion progress, October 2026. This article was prepared by the expert team at Global Infinity Labs, drawing on insights and data provided through its MCP data services. About Global Infinity Labs Global Infinity Labs is an AI innovation company specializing in MCP data services. Email: support@globalinfinitylabs.com To share industry insights, report corrections, or submit story tips: editors@globalinfinitylabs.com or https://www.globalinfinitylabs.com/submit.
More
Southeast Asia Wants Electric Delivery Fleets. The Math Keeps Changing. SeaPRwire

Southeast Asia Wants Electric Delivery Fleets. The Math Keeps Changing.

Two-wheelers, subsidies and battery subscriptions across Indonesia, Thailand and Vietnam in 2026: the real cost of electrifying the region’s riders Somewhere in Bekasi or Surabaya this morning, a food delivery rider is looking at an electric motorcycle and doing arithmetic that no policy paper has finished for him. His petrol bike refuels in two minutes. The electric alternative needs four hours on a charger, unless he can reach one of the battery-swap cabinets springing up at convenience stores. The bike itself costs more than the petrol equivalent, unless the battery is rented separately, which it usually now is. And the government subsidy that was supposed to close the gap has a habit of arriving, running out, and disappearing. That arithmetic is the actual bottleneck of Southeast Asia’s electric mobility story. The region’s headline numbers look unstoppable: Southeast Asia sold roughly 800,000 electric vehicles in 2025 and is forecast to pass 1.2 million in 2026 (industry estimates, 2026). But the vehicles that matter most to daily life, the two-wheelers that carry food, parcels and people, sit at the bottom of a broken incentive chain. Indonesia has 139 million registered motorcycles and only around 236,000 of them are electric (Hello Power, 2026). The gap is not a technology problem. It is a cash-flow problem, and the three ways out of it cost very different things. The rider’s problem is time, not range A delivery rider in Southeast Asia does not earn by the kilometre. He earns by the completed trip, and every idle hour is unpaid. That is why the charging question dominates everything else. “A food delivery rider in Bekasi or Surabaya cannot wait four hours for an electric motorcycle battery to charge, while conventional bikes refuel in two minutes,” as one 2026 market analysis put it (Hello Power, 2026). Battery swapping, where a depleted pack is exchanged for a charged one in under a minute, has therefore become the dominant infrastructure approach for professional riders in the region. Underneath the time problem sits a cost problem. The lithium-ion battery typically accounts for 30 to 40 percent of an electric motorcycle’s total cost (Tribun Digital, September 2026). A rider buying a bike is really buying a bike plus a depreciating chemistry set, and resale values for used batteries remain uncertain. This is why manufacturers have spent 2025 and 2026 rebuilding their pricing around battery rental, and why the purchase price alone no longer tells a rider what he will actually pay. The third constraint is income volatility. Indonesia’s own driver surveys, filed during the June 2026 protest wave, recorded drivers working 12 to 18 hours a day with monthly income around 52.6 percent of Jakarta’s minimum wage of IDR 5.7 million (SPAI statement, June 2026). A rider at half the minimum wage cannot absorb a wrong vehicle decision. Whatever path he takes has to survive a bad month. Why the usual fixes do not hold The obvious fix, a purchase subsidy, has been tried and has a track record of backfiring. Indonesia’s first electric motorcycle subsidy, IDR 7 million per unit under Permenperin No. 21/2023, delivered 62,541 units before being wound down in 2025 (Koran Manado, May 2026). When the subsidy stopped, electric motorcycle sales dropped by roughly half in the first half of the year (Hello Power, 2026). The lesson the industry drew was blunt: demand that only exists when subsidies are available is not really demand. The 2026 program restarted from June with a smaller allocation: IDR 5 million per unit, roughly USD 289, against a quota of 100,000 electric motorcycles, with 100,000 electric cars getting VAT relief of 40 to 100 percent depending on battery type (Finance Minister Purbaya Yudhi Sadewa, May 2026, via ANTARA). The industry association Periklindo immediately asked the government to prioritise ride-hailing drivers over private first buyers, with its chairman Moeldoko arguing that transport corporations such as Gojek should be first in the queue so the quota serves people who ride for a living (Koran Manado, 2026-05-22). Infrastructure is the second fix that lags. Indonesia plans 52,000 to 67,000 battery swap stations nationwide by 2030 and currently operates around 2,000 (Hello Power, 2026). Thailand, the region’s charging leader for four-wheelers, counts 3,720 public charging stations with 11,622 points (Borderless, 2026), nearly all of them built for cars rather than the two-wheelers that dominate Thai traffic. Vietnam has taken a different route: VinFast’s charging arm V-Green operates around 150,000 charging ports, an enormous number that mostly serves one manufacturer’s ecosystem (Borderless, 2026). The region had about 425,000 public chargers of all kinds in the first quarter of 2026 (industry statistics, 2026), and almost none of that network was designed around a delivery rider’s working night. Meanwhile the platforms themselves keep shifting the rider’s income side. Indonesia’s ride-hailing commission settlement, announced on May 1, 2026 after protest waves, cut the platform commission cap from 20 percent to 8 percent and lifted the driver revenue share from 80 to 92 percent, executed from July 1 and formalised in PR 27/2026 issued October 8 (Tech in Asia; ANTARA). That reform raises the rider’s take-home per trip, which is exactly what makes an electrification decision thinkable. It also arrives years late and could be revised again. The income side finally moved, quietly It is worth putting numbers on that income shift, because electrification decisions are made against monthly cash flow, not against sentiment. Grab’s first quarter of 2026, the most recent full set of published platform financials for the region, showed revenue of USD 955 million, up 24 percent year on year, with group monthly transacting users at 51.6 million and on-demand GMV of USD 6.13 billion (Grab investor report, 2026). The company paid USD 650 million in total incentives that quarter, 10.5 percent of on-demand GMV, money that functions as a temporary wage subsidy for riders and a temporary discount for passengers. Two details in that quarter’s data frame the electrification question. First, two-wheeler rides were only about 6 percent of Grab’s mobility GMV in Indonesia in the quarter, per its CFO commentary (Grab, 2026), meaning the region’s largest platform still earns most of its ride revenue from cars even in the most two-wheeled big market on earth. Second, the incentive load means a rider’s effective income includes platform money that can be withdrawn in a cost-cutting quarter. A rider who buys an electric motorcycle against a take-home that assumes USD 650 million a quarter of incentives is buying against a subsidy he does not control. The commission reform, 8 percent capped and 92 percent revenue share, is the first piece of the income stack written into regulation rather than platform choice, which is why Indonesia’s driver associations treat PR 27/2026 as the more durable change. What the neighbours did instead Indonesia is not the only laboratory, and its two largest neighbours made opposite bets worth checking against the same rider math. Vietnam bet on a vertically integrated manufacturer. Battery electric vehicles took 35.3 percent of its new car market in the first half of 2026, with 115,986 BEVs sold, and VinFast held roughly 40 percent of the domestic market (VnExpress; ICCT, 2026). On two wheels, VinFast’s charging arm V-Green runs about 150,000 charging ports, an unmatched number that mostly serves one ecosystem (Borderless, 2026). The registration-fee exemption for BEVs runs to 2030 under the current schedule, and the special consumption tax on BEVs sits at 3 percent. A Vietnamese rider entering this system buys into a single company’s hardware, charger and service network, with 400 workshops nationally. Thailand bet on charging infrastructure and imports. The country counted 3,720 public charging stations with 11,622 points in 2026, and Chinese brands held 91 percent of its battery electric vehicle market (Borderless; industry statistics, 2026). Thai BEV sales reached 104,418 units in the first half of 2026. The network is genuinely dense by regional standards, but it was built around four-wheel charging, and motorcycle taxi conversions, a mainstay of Bangkok’s streets, still wait for a two-wheeler analogue. Across the six largest ASEAN markets, electrified vehicles took about 32 percent of vehicle sales in the first half of 2026, with Chinese brands at 16 percent of total sales and 47 percent of the electrified slice (PwC, 2026). The trend lines differ by country, but every one of them routes through the same rider arithmetic this article started with: a vehicle the rider can afford, energy he can get in under two minutes, and an income he can schedule around. Path one: buy the bike, rent the battery The first executable path is the manufacturer battery subscription, known in the market as Battery-as-a-Service. The rider buys the chassis at a reduced price and rents the battery monthly. As of September 2026, Indonesian subscription tariffs run from Rp84,000 to Rp250,000 per month depending on brand and battery count (Tribun Digital, September 2026). The price points are public and worth laying out, because the spread between them is the market sorting itself. VinFast, the lowest entry: Rp84,000 per month for a single-battery Evo, Feliz II or Viper, Rp144,000 for dual-battery configurations. Polytron: Rp125,000 per month for the Fox 200, Rp200,000 for the Fox 500 and Fox R, with a contractual promise to replace the battery if its state of health drops below 85 percent. ALVA, the premium tier: Rp150,000 for a single-battery N3, Rp250,000 for dual-battery N3 or CERVO, taxes included. The 2026 subsidy of IDR 5 million stacks on top of this path, lowering the entry price further. The costs are a monthly obligation that never ends, a lock-in to one manufacturer’s battery format, and exposure to subsidy timing, since the quota of 100,000 units is phased and the money has not always been final. This path suits a rider with steady trip volume in a city where the manufacturer’s charging network exists, and it is the default choice for most first-time buyers in 2026. Path two: join the swap network The second path treats the battery as infrastructure. The rider pays per swap, or a subscription to a network, and never owns a battery at all. Indonesia is the regional centre of this model and holds 37.5 percent of the Southeast Asia battery swapping market (Hello Power, 2026). SWAP Energi, the largest locally built operator, runs more than 1,500 stations concentrated at Circle K outlets, Alfamart stores and warungs along delivery corridors in Jakarta, Surabaya, Bali and Bandung, and holds fleet partnerships with Grab and Lazada that anchor utilisation before a station even opens (Hello Power, 2026). Electrum, the joint venture between GoTo and TBS Energi built on Gogoro’s platform, deployed more than 370 swap stations supporting over 6,000 electric motorcycles across Greater Jakarta by 2025, with a captive demand base in Gojek’s driver network (Hello Power, 2026). Even Pertamina Patra Niaga, the state oil company, has installed 46 swap units at its petrol stations in Greater Jakarta, a hedge for the day fuel demand declines. The costs here are per-swap fees that scale with usage, dependence on station density, and a real geographic wall: the network barely exists outside the major urban corridors. A rider in a secondary city cannot buy his way into this path at any price. For a high-volume rider inside the covered corridors, though, the swap model is the only one whose refuelling time competes with a petrol station, and the platform partnerships mean the demand side is contracted rather than hoped for. Path three: convert the bike you already have, or wait The third path is the least glamorous and possibly the most honest: retrofit the existing petrol motorcycle. Indonesia’s energy ministry has said conversion costs have fallen to the IDR 5 to 6 million range, roughly USD 300 to 360, as the technology matures (Indonesia Daily, 2026). The government has floated conversion subsidies modelled on its earlier schemes. The economics are attractive on paper: a conversion costs about the same as the 2026 purchase subsidy itself, and it preserves an asset the rider already owns. The risks are equally real. Converted bikes carry no manufacturer battery warranty, no state-of-health guarantee, and an uncertain resale story. Quality varies by workshop. A rider who converts cheaply and gets a degraded battery has converted nothing. The variant of this path that most riders actually take is a fourth option the government does not subsidise: do nothing, keep the petrol bike, and stay exposed to fuel prices, which is precisely the exposure the current subsidy wave exists to relieve, given that the 2026 program was triggered by oil prices elevated by Middle East conflict (Finance Ministry, May 2026). Choosing by the rider’s own math None of the three paths dominates. The right choice depends on three variables a rider can actually observe: his daily distance, his city’s swap density, and his tolerance for a permanent monthly bill. Table 1: Three electrification paths for Southeast Asian two-wheeler riders, 2026 Path Upfront cost Ongoing cost Main risk Fits whom Buy chassis, rent battery (BaaS) Reduced purchase price minus IDR 5M subsidy Rp84,000 to Rp250,000 per month Perpetual subscription and brand lock-in Steady-volume riders in cities with the brand’s network Swap network membership Low or none for the battery Per-swap fees scaling with usage Station density outside major corridors High-volume riders in Jakarta, Surabaya, Bali, Bandung Retrofit existing bike IDR 5-6 million conversion Charging or swap fees only No warranty, uncertain battery quality Rural riders and owners of recent petrol bikes Source: Tribun Digital , Hello Power The honest answer to “which path wins” is that the region has not built the conditions for a single winner. Vietnam is betting on a vertically integrated ecosystem around one manufacturer. Thailand is betting on charging points sized for cars. Indonesia is betting on swap infrastructure with platform-contracted demand, and it is the only country of the three whose policy explicitly discusses routing subsidies to ride-hailing fleets first. Where the delivery volume is densest, the swap path is already winning. Where it is not, the BaaS subscription is the only path with a price tag a half-minimum-wage income can schedule around, which is why it dominates first purchases even among riders who secretly envy the swap cabinets. What should a rider do this quarter? If he rides more than about 60 kilometres a day inside a covered swap corridor, join the network. If he rides less and his city has no swaps, take the subsidised BaaS package and treat the monthly fee as his fuel budget renamed. If he is rural, the cheapest defensible move in 2026 is still to wait, because the conversion market’s battery quality has not yet earned the IDR 5 million it would cost him. The arithmetic keeps changing, but for once it is changing in the rider’s favour, one repealed and reissued subsidy at a time. Data basis The purchase and subscription prices in this article are manufacturer-published Indonesian market tariffs as of September 2026, taken from market round-ups (Tribun Digital) and a 2026 infrastructure analysis (Hello Power), not from field verification. The subsidy figures, IDR 7 million for the 2023-2024 scheme that delivered 62,541 units and IDR 5 million for the 2026 scheme with a 100,000-unit quota starting June 2026, come from Indonesia’s finance ministry statements of May 2026 as carried by ANTARA and Indonesian press. Driver income figures come from the SPAI driver association’s June 2026 statement and are advocacy numbers, not government statistics. Swap station counts are operator claims as of 2025-2026 and cannot be verified against an official register. Sample size for rider economics is effectively one country, Indonesia, because that is where the published data lives; Thailand and Vietnam appear through their charging network counts only. Where two sources disagreed on a figure, the more conservative one is used. Data as of November 2026. Sources Hello Power , Battery swapping for electric motorcycles in Indonesia, 2026 Tribun Digital , Battery subscription guide for Indonesian electric motorcycles, September 2026 Koran Manado , Periklindo asks priority for ojol drivers, May 22, 2026 (in Indonesian) Medcom.id , Indonesia EV incentives to curb fuel imports, May 12, 2026 Indonesia Daily , Indonesia continues electric motorcycle subsidy in 2026 (in Chinese) ANTARA via Xinhua , Indonesia plans EV subsidies from June 2026, May 13, 2026 About Global Infinity Labs Global Infinity Labs is an AI innovation company specializing in MCP data services. This article was prepared by our expert team using insights and data provided through our MCP platforms. Contact: support@globalinfinitylabs.com
More
If Family Premiums Rose 5%, Why Did Workers’ Bills Jump 9%? SeaPRwire

If Family Premiums Rose 5%, Why Did Workers’ Bills Jump 9%?

An audit of the 2026 US employer health cost numbers: KFF’s $28,303 premium, Milliman’s $37,824 cost of care and the 7.3% national spending growth that are not the same thing The most quoted number in American employer health insurance this autumn is 5 percent. That is the average increase in the family premium for employer-sponsored coverage in 2026, according to the Kaiser Family Foundation’s 28th annual Employer Health Benefits Survey. The second most quoted number is 9 percent. That is the increase in what covered workers themselves paid toward family coverage this year, up $621 to an average $7,471. Both figures come from the same survey of the same 1,937 employers. A reader encountering them in separate stories could reasonably conclude that US health costs either grew modestly or grew alarmingly. The answer is that neither summary is wrong, and neither is the whole story, because the two numbers sit on different denominators. This article does what an auditor would do with the 2026 batch of health cost figures: quote each claim in its publisher’s own words, take each denominator apart, cross-check the numbers against each other, and identify which pairs of statements genuinely cannot be reconciled and which merely sound contradictory. The finding, in short: every major number survives its own definition, and the public confusion is manufactured by the definition-hopping that happens between headlines. The claims, as published Start with the claims as published, before anyone translates them. KFF’s news release headline: “Workers Now Pay $7,471 on Average Annually for Family Health Coverage, Up 9% or $621 from 2025” (KFF, October 22, 2026). The survey body adds: annual premiums for employer-sponsored family coverage reached $28,303 in 2026, 5 percent higher than 2025; the average single premium rose 5 percent to $9,804 in 2026; workers’ wages rose 3.7 percent and inflation ran 2.7 percent over the same twelve months. KFF president Drew Altman connected the numbers to politics: “We have been hearing a lot about Affordable Care Act and Medicaid cuts. Now add a 9% increase in what workers with employer coverage paid toward family premiums in 2026 to the reasons voters are so concerned about their health care costs” (Healthcare Finance News, October 2026). The Milliman Medical Index, published mid-year, states that the total cost of care for a typical American family of four covered by an average employer PPO plan is projected at $37,824 in 2026, up 7.2 percent, or $8,460 per person, up 7.9 percent (Milliman, 2026). The federal baseline comes from the National Health Expenditure accounts: US health spending grew 7.3 percent in 2025 and reached 18.4 percent of GDP, with 91.7 percent of the population carrying insurance coverage (CMS Office of the Actuary via Health Affairs, 2026). Four publishers, four numbers, four different things. Now the denominators. Taking the denominators apart The KFF premium is a price, not a cost. The $28,303 family premium is the total annual price of an employer-sponsored insurance policy, employer share plus worker share combined, averaged over covered workers at non-federal public and private firms with ten or more employees, from 1,937 interviews conducted January through July 2026 (KFF, 2026). It excludes the under-65 population without employer coverage, deductibles, and out-of-pocket spending. KFF covers almost 157 million people under 65 through this lens, which is the survey’s own denominator for “employer-sponsored insurance”. The worker contribution is a slice of that price. The $7,471 is the worker’s share of the same family premium, about 26 percent of $28,303. When news stories say “workers’ bills jumped 9 percent,” the denominator is this slice, not the whole premium. The 5 percent premium growth and the 9 percent contribution growth coexist because employers shifted the split: among large firms, 44 percent increased employees’ share of premiums and 21 percent raised deductibles or other cost-sharing in response to projected cost increases (KFF, 2026). The Milliman figure is a cost, not a price. The $37,824 measures total cost of care for a family of four in an average employer PPO plan, including what employees pay out of pocket, projected actuarially (Milliman, 2026). It is a family-of-four unit, not a covered-worker unit, and a PPO plan, not the all-plan average. Milliman’s number being 33 percent larger than KFF’s premium is not a contradiction; it is a larger bucket. The national growth rate is a different economy entirely. The 7.3 percent NHE growth in 2025 includes Medicare, Medicaid, ACA marketplaces, out-of-pocket spending, and every premium dollar in the system, measured against the whole economy (CMS OACT, 2026). A national aggregate cannot be read as the trend facing a specific worker at a specific firm. In 2025 the large-group market decelerated while marketplace and government programs accelerated, so the national rate and the employer rate legitimately diverge. Cross-checks between the numbers Three internal consistency checks, each with a defined denominator. Check one: the five-year stack. Over the last five years, KFF’s family premium rose 28 percent, workers’ wages 26.9 percent, and inflation 24.4 percent (KFF, 2026). The premium has outpaced both wages and prices over the period, but only by 1 to 4 percentage points cumulative, which is a modest gap. This makes the single-year 9 percent worker contribution jump look less like a trend and more like an adjustment year after two unusually soft contribution years. Check two: the small-firm split. Workers at firms with 10 to 199 employees paid an average $9,687 toward family coverage in 2026, versus $6,840 at large firms, a gap of $2,847 in the 2026 survey year (KFF, 2026). Nearly a third of small-firm workers, 32 percent, were enrolled in plans requiring them to pay more than half the family premium, versus 6 percent at large firms. Single-coverage deductibles averaged $2,624 at small firms against $1,772 at large ones. The “average worker” the headlines describe is a blend of two very different populations. Check three: the forward-looking filings. Preliminary Marsh survey findings, fielded in October 2026, put employer expectations for 2027 health benefit cost growth at 8.2 percent per employee after planned changes, or 11 percent without them (Healthcare Finance News, 2026). Small-group insurers sought a median 14 percent premium increase for 2027 in their rate filings. These are proposals and expectations, not premiums paid, and they belong in a different column from KFF’s realized 5 percent. If KFF’s 2027 survey lands near 8 percent rather than 5, the filings will have been the better leading indicator. The employers’ playbook, 2026 edition The survey also records what employers did when they saw the cost curve coming, and the answers explain the 9 percent worker contribution number mechanically rather than mysteriously. Among large firms offering health benefits, 44 percent increased employees’ share of premiums in response to projected cost increases. 21 percent raised deductibles or other cost-sharing for most services. 16 percent changed insurers or plan administrators, and 10 percent added utilization management requirements such as prior authorization. Among firms with 5,000 or more employees, 66 percent had already raised worker contributions in anticipation of cost growth (KFF, 2026). Two structural details sharpen the picture. First, level-funded plans, a self-funded design with stop-loss insurance that had mostly been a large-employer instrument, now cover 38 percent of covered workers at small firms, moving small employers onto a claims-exposure model many of them do not fully understand (KFF, 2026). Second, the GLP-1 stalemate held: 18 percent of large firms covered the drugs for weight loss in 2026, unchanged from 2025 and 2024, 12 percent stopped covering them within the past year, and among firms that do cover them, 29 percent report significant prescription spending impact (KFF, 2026). A separate Business Group on Health survey found large-employer GLP-1 weight coverage falling from 72 percent to 60 percent between 2025 and 2026 (Insurance Business, October 2026). The 2026 worker contribution increase therefore lands in a year when employers were actively refusing one new cost category while shifting old ones to workers. The timing matters as much as the direction. The 44 percent of large firms that shifted premium share did so during 2026 plan years, meaning the change hit workers in the same twelve months that open enrollment messaging still described stable cost-sharing. And the 21 percent that raised deductibles did so while the average deductible was already 34 percent above its 2016 level, so the marginal worker at a small firm absorbed a contribution increase in 2026 on top of a decade of cost-sharing increases that preceded it. Table 2: The 2026 employer response to projected cost increases, large firms Action taken Share of large firms Ref Increased employees’ share of premiums 44% KFF 2026 EHBS Raised deductibles or cost-sharing 21% KFF 2026 EHBS Changed insurers or administrators 16% KFF 2026 EHBS Added utilization management 10% KFF 2026 EHBS Source: KFF What the average hides: plan types, regions, deductibles The single “average family premium” also blends plan designs and geographies that behave differently, and the differences matter for anyone renewing a policy in 2027. By plan type, workers enrolled in PPO plans carry above-average premiums, $10,272 for single coverage and $29,522 for family in 2026, while workers in HSA-qualified plans sit below the overall average at $8,989 and $26,450 (KFF, 2026). The spread is roughly 15 percent between the most and least expensive common designs, which means a firm changing its plan mix can report a slower premium increase without any worker’s actual plan getting cheaper. By region, the Northeast ran the highest premiums in 2026 at $10,806 single and $31,378 family, against the South’s $9,272 and $26,702 (KFF, 2026). A 5 percent national increase is consistent with near-zero growth in one region and 8 percent in another. Deductibles add a second layer the premium discussion skips. The average general annual deductible for single coverage among workers with one was $1,955 in 2026, statistically flat against 2025, but that calm headline sits on a steeper history: the same figure has risen 18 percent over five years and 34 percent over ten (KFF, 2026). More than half of small-firm workers, 55 percent, are now enrolled in a plan with a single-coverage deductible of $2,000 or more, against 33 percent at large firms. For that population the real annual exposure is contribution plus deductible, a number no headline prints. The 9 percent in historical context The worker contribution increase deserves its own denominator question: 9 percent of what baseline, and how unusual is it? KFF’s own reading is that the worker share of premiums held relatively stable for decades before 2026, which is why the survey flagged this year’s shift as a potential affordability turning point (KFF, 2026). A 9 percent rise on a $6,850 base, roughly where the 2025 family contribution sat, is $621. The same percentage applied five years ago would have been closer to $450 on the smaller base. The rate is new; the dollars are the accumulated base catching up. Who absorbs the increase also changed in 2026. At firms with 5,000 or more employees, 66 percent said they had already raised worker contributions in response to anticipated cost growth (KFF, 2026). At small firms, the pressure shows up in structure rather than announcements: nearly a third of their covered workers sit in plans requiring them to pay more than half the family premium, and 38 percent of small-firm workers are in level-funded designs where a bad claims year can flow back toward the employer, and eventually the workforce, at renewal. Where the numbers genuinely cannot be reconciled After the denominators are separated, one real inconsistency remains, and it is worth naming precisely. The 5 percent premium growth and the 9 percent worker contribution growth cannot both describe “what happened to the cost of family coverage” for a worker at a small firm, because KFF’s own small-firm data shows the worker share moving much faster than the premium. The reconciliation exists, employers absorbed less of the increase, but the average figures conceal it. A worker at a firm with 5,000 or more employees saw her contribution rise toward a $28,592 average premium with a 66 percent likelihood her employer raised her share in anticipation of cost growth (KFF, 2026). A worker at a 40-person firm faced a $27,288 premium and a one-in-three chance of paying more than half of it. The single-number story breaks exactly where the stakes are highest. A second, softer tension: NHE grew 7.3 percent in 2025 while employer premiums grew 5 percent in 2026. These are different years, different baskets, and different price bases, so no mathematical contradiction exists. But any story that quotes the national 7.3 percent next to the employer 5 percent as if they described the same phenomenon is committing a denominator error, and the pattern repeats in both directions across the 2026 coverage. Verdict The 2026 numbers, correctly read, say this. The sticker price of employer family coverage rose 5 percent to $28,303. Workers’ share of that sticker rose 9 percent because employers pushed cost-sharing down the hill, especially at small firms, where nearly a third of covered workers now pay more than half the premium. The underlying cost of care, measured actuarially, runs 7.2 percent. The system as a whole grew 7.3 percent in 2025 against an economy where wages rose 3.7 percent. Nothing in that set is fraudulent, and nothing in it is a simple sentence. For the audit, one rule carries the verdict. Quote any of these numbers without its denominator and you will mislead a reader, including, most weeks, yourself. The figure most likely to be misused in 2027 arguments is KFF’s 5 percent, because it is the lowest number in the set and the only one that describes a price rather than a bill. For a benefits manager renewing in January 2027, the practical translation is unglamorous. Expect an 8 percent cost trend before plan changes, per the Marsh expectations, which means the 2026 playbook of shifting share to workers has one more round left in it before the affordability question moves from the benefits team to the compensation committee. For a journalist quoting the 2026 batch, the rule is simpler: name the denominator or drop the number. And for the 32 percent of small-firm workers already paying more than half their family premium, no denominator hides the fact that their bill grew twice as fast as the premium everyone is quoting. Methodology This audit used four primary sources: KFF’s 2026 Employer Health Benefits Survey (28th annual, 1,937 employer interviews, fielded January to July 2026, non-federal public and private firms with 10 or more workers), the Milliman Medical Index 2026 (actuarial projection of total cost of care for a family of four in an average employer PPO plan), the CMS Office of the Actuary National Health Expenditure data for 2025 as published via Health Affairs, and secondary trade coverage (Healthcare Finance News) for the Marsh 2027 expectations and small-group rate filings. For each figure, we recorded the denominator as defined by its publisher before comparing numbers. Time frames are 2026 realized premiums, 2025 national expenditure, and 2026-2027 projections. We did not reweight or adjust any source’s figures, did not mix projection with realized values, and treated the small-group 14 percent median filing as a proposed rate rather than a cost. Definition of “increase” follows each publisher’s own year-over-year comparison. Data as of November 2026. Limitations: KFF’s survey is employer-reported and excludes federal employers and firms under 10 workers. Milliman’s index is a projection, not an invoice. The Marsh figure is preliminary. No source here measures the uninsured population’s costs. Sources KFF , 2026 Employer Health Benefits Survey, October 22, 2026 Healthcare Finance News , Family health premiums top $28,000 as worker contributions rise 9%, October 2026 Health Affairs , CMS OACT national health expenditure projections, 2026 CFH Insurance Center , Milliman 2026 Medical Index comparison, 2026 Insurance Business , Small-firm workers pay $2,823 more for family coverage, October 2026 About Global Infinity Labs Global Infinity Labs is an AI innovation company specializing in MCP data services. This article was prepared by our expert team using insights and data provided through our MCP platforms. Contact: support@globalinfinitylabs.com
More