aiminute. ← All AI news
New Models 2026-08-14

Apple built its own model for China — the first foreign company Beijing has let do it

Apple built its own model for China — the first foreign company Beijing has let do it

Reuters reported on 14 August, citing three people familiar with the matter, that Apple has trained a large language model specifically for the Chinese market with technical support from Alibaba. That is a departure from the plan to run Apple Intelligence in China on a domestic partner's model alone. The Cyberspace Administration of China completed registration of Apple's generative AI service last month, which according to the report makes Apple the first foreign company cleared by Beijing to offer a proprietary AI model on the mainland. Alibaba's Qwen stays part of the arrangement and Baidu has worked on Chinese-language features. The rollout is expected within months, possibly alongside September's iPhone launch. Neither Apple nor Alibaba has commented.

Why it mattersThe working assumption for two years was that a foreign company could not run its own model in China — it had to rent a domestic one and accept what that meant for control over its own product. If this registration holds, Apple has found a third path: its own model, trained with a local partner, cleared by the local regulator. Every US firm currently locked out of that market will read it as a template. One caveat worth keeping: this rests on a single wire report built on unnamed sources, and neither company has confirmed it.

✓ Verified · 4 sources

WhatsApp X Telegram
Read in the app — free, in 9 languages

Related stories

DeepSeek's cheap workhorse can now see — and on agent tasks that need eyes it says it is close to Anthropic's best
2026-08-21
Show the robot once — three to twelve seconds — and it gets the job right 59 times out of 100 with no training at all
2026-08-21
The model invents its own tasks, builds the rig to test them, then trains on the results — and DeepReinforce gave the weights away
2026-08-20
A free model that fits on one graphics card scores 52 — then spends 22,276 thinking tokens drawing a picture
2026-08-18
Zhipu did not build a new model. It kept training the old one — and says coding got 50 percent better.
2026-08-17