01.AI winds down its open model platform

China's AI shakeout produced its clearest retreat of the week: Kai-Fu Lee's 01.AI is winding down its developer platform, while Chinese robot makers extended their shipment dominance and the country's model price war finally started reversing.

01.AI, the model startup founded by Kai-Fu Lee, will shut down its open model platform, stopping online demos, API calls and recharge services for developers, according to reports in DoNews and Sina. Once counted among China's "six little tigers" of AI startups, 01.AI told users the platform will gradually wind those services down; the company has not publicly commented. It is the latest step in a long retreat: 01.AI abandoned trillion-parameter frontier pretraining in 2025, saw its pretraining and infrastructure teams fold into Alibaba's, and pivoted to enterprise deployment through its Wanzhi platform. Now the developer-facing API business is being ceded too. The message to the market is blunt — independent labs are exiting the commodity model-API race one by one, leaving DeepSeek and Alibaba's Qwen to own developer mindshare, and developers still on Yi APIs are left planning a migration.

Chinese manufacturers shipped more than 97% of the world's humanoid robots in the first half of 2026 — roughly 19,100 units, nearly four times the 5,100 shipped a year earlier, according to Bloomberg's Saritha Rai. The numbers, carried by Techmeme, quantify a dominance the market has watched build all year: Unitree's landmark mainland IPO and a stream of home-robot funding rounds kept the sector in headlines (we covered Unitree's $9 billion IPO — Unitree prices $9B humanoid IPO, first in mainland China — and the ex-Huawei team's $70 million raise — Ex-Huawei team raises $70M for fold-up home robot). The practical consequence cuts both ways: Western buyers get cheaper robots, but the supply chain, training data and scale economics are consolidating in one country.

DeepSeek told developers it plans a "significant" across-the-board API price increase, and Alibaba is weighing a revenue-sharing license for its next flagship Qwen3.8-Max — the clearest sign yet that China's two-year model price war is reversing. DeepSeek's API documentation now flags an upcoming peak/off-peak pricing scheme that would double prices during Beijing rush hours (9:00–12:00 and 14:00–18:00), on top of current V4-Flash rates of 1 yuan per million input tokens and 2 yuan per million output tokens. Reuters, cited by TMTPost, reported that Alibaba's new license would keep Qwen3.8-Max's weights open but require large commercial users to share revenue, with the split ratio and license text still under discussion. The leverage is building: Chinese models have now led global API call volume for 15 straight weeks, with the official DeepSeek-V4-Flash release topping the latest chart — when everyone's already using your model, charging more starts to look safe.

What to watch: whether other Chinese labs follow 01.AI out of the developer-platform business — and how big DeepSeek's "significant" increase actually lands.

When a price war ends and a platform shuts down in the same week, who's left standing in China's model race? Tell us in the comments.

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

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

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

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

Read more

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

The Week in AI — October 5–11, 2026

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

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