Hubei opens a state-run token exchange for 200+ AI models

Share
Hubei opens a state-run token exchange for 200+ AI models

China's provinces are turning model access into a utility: Hubei just opened a government-linked marketplace where companies buy tokens like phone credit, while Apple's first AI home device moves into employees' houses and new data shows how far AI has actually reached China's factory floors.

Hubei Big Data Group launched a Token Trading Center on September 20 at the province's computing industry ecosystem conference, a single platform where an enterprise signs up once and can call any of the 200-plus model services already listed. The center claims 326 ecosystem partners and estimates a 15% to 30% cut in what companies spend on models, achieved through intelligent routing, semantic caching across identical prompts, and failover between providers. Billing borrows the telecom playbook: customers top up a token account in packages and spend it across different vendors, rather than opening a separate relationship with each lab.

The architecture is the interesting part, because it is a governance product as much as a shopping one. Four layers do the work — aggregation (one entry point, filterable by model type, vendor and capability), a gateway that handles key custody, content filtering and full-chain audit under unified protocols and single sign-on, a scheduling layer that picks the cheapest competent model per call, and a dashboard that breaks token spend down by organization, department, project or individual with budgets, quotas and anomaly detection. That last layer is what a finance department actually buys: the end-of-month surprise AI invoice becomes a live meter. The platform builds on Hubei's national compute-interconnection regional node and its provincial AI public-service platform, positioning the central province as the settlement layer between western power and compute and eastern demand. China's National Data Administration wrote "token" into a national policy document for the first time this year and floated token trading as a new transaction model; provincial five-year plans are now competing to write the settlement rules. We covered the demand side of this market in September — Zhipu's API revenue jumps 27x as China's AI labs prove the token economy.

Why it matters: when a state intermediary owns metering, routing and audit, the labs stop competing on price alone and start competing to be the model the router picks. The margin migrates from per-token pricing to the switching layer — and whoever holds the switch sets the terms.


Apple's long-delayed AI home hub is now being tested in employees' homes, with a launch as soon as October, according to Mark Gurman's Power On newsletter. The device, internally tracked as J490 and pitched inside Apple as an "intelligent personal hub," runs a Siri-AI-centric operating system derived from the Apple TV line rather than iOS, with a roughly seven-inch screen and both desktop and wall-mount variants. At-home employee testing is the final stage before a product ships, and the same newsletter reported Apple cutting staff in its Fitness+ division as it reassesses that service's future. Apple has not announced the device or a date; the window Gurman describes runs from October into early 2027.


Manufacturing AI reached 34.2% penetration across China's key industrial scenarios in the first half of 2026, according to a report released at the World Manufacturing Convention in Hefei on September 20. The study, Manufacturing Digital-Intelligent Transformation Capability Level (2026), comes from the state-run National Industrial Information Security Development and Research Center, and puts the national integration level for digitalization and industrialization at 67.1. Read the two numbers together and the picture is a country in the messy middle of adoption: a third of priority factory use cases have a model in the loop, and the digital plumbing underneath is only two-thirds built.

What to watch: whether Hubei's exchange signs model vendors on exclusive terms, and whether Apple's hub ships with Siri AI features that are still waitlisted in the US and blocked in the EU.

Would you rather buy AI capacity from a state-run marketplace or contract each lab directly? Tell us in the comments.

Read more

Akhetonics says its all-optical CPU reaches a customer in 2026

Akhetonics says its all-optical CPU reaches a customer in 2026

Light-based computing keeps promising more than it delivers — but one Munich startup has just put a date on its bet, and the interview laying it out is doing the rounds on Hacker News this week. Akhetonics says it will deploy its first commercial machine with a major customer by the end of 2026, with several more planned for 2027. The company, founded by Michael Kissner and Leonardo Del Bino, is building a computer where data enters as light, is switched as light, and circulates through memory

The Week in AI — October 5–11, 2026

The Week in AI — October 5–11, 2026

Every big claim this week turned out to rest on fine print more interesting than the headline: revenue only the company reporting it can define, safety tests sandboxed while the product keeps the web, and a Pentagon phase-out nobody would confirm until reporters kept asking. The week's top 5 1. OpenAI's revenue was $20 billion below the numbers everyone quoted — and the gap was definitional. The Financial Times reported Thursday that OpenAI's annualized revenue runs roughly $20 billion unde

Drone strike shuts a third Yandex data center, taking YandexGPT offline

Drone strike shuts a third Yandex data center, taking YandexGPT offline

Russia's largest tech company is learning what the AI era's infrastructure war looks like from the receiving end — three data centers in four days, and with them much of the cloud layer Russian businesses run on. A Ukrainian drone strike knocked out Yandex's data center in Vladimir early Sunday morning, the third of the company's facilities hit since October 8. The site — reported at roughly 50 MW and designed for about 2,880 server racks — stopped operating completely after the attack, Yandex

Agent teams cost up to 5x more, barely score higher

Agent teams cost up to 5x more, barely score higher

The multi-agent hype train hit a benchmark this weekend — and the grid and the trucking regulators had quiet weeks of their own. Vals AI put agent teams head-to-head with single agents on its Vibe Code Bench, and the teams cost between 1.8 and 5.1 times more for almost no extra quality. The evals company ran GPT-6 Sol and Claude Opus 5.5 solo and in teams across 50 apps at two reasoning efforts; out of four comparisons, only one was statistically significant — Sol at medium effort, where the t