Zhipu AI (智谱) reports its MaaS open platform now has nearly 7 million registered API users, roughly...
The AMW Read
Update on Zhipu's scale-up with explicit compute and silicon shift, but not resolving a debate or new entrant.
Zhipu AI (智谱) reports its MaaS open platform now has nearly 7 million registered API users, roughly 2 million more than in early July, according to LatePost via TechNode. To handle rising inference demand, the company says it has newly activated more than 50,000 domestically developed AI chips. Zhipu also reopened purchases of its previously restricted Coding Plan on July 31 after raising the price.
The user growth signals strong enterprise and developer adoption of Zhipu's models—GLM series—at a time when China's foundation-model market is intensifying. The chip activation underscores a strategic pivot to domestic silicon, likely to reduce dependence on Nvidia GPUs amidst ongoing export controls. This dual move—scaling API users while expanding compute capacity with local chips—positions Zhipu as a key player in China's sovereign AI infrastructure buildout. The concurrent coding-plan price increase suggests growing confidence in demand for its developer tools.
For builders, the near-7M API user base means a large, active ecosystem for GLM models, with potential for broader tooling and integration opportunities. For investors, this signals that China's AI labs are aggressively scaling both user reach and compute independence. The shift to domestic chips could also open a market for Chinese AI accelerators, making them a closer-to-commercial reality.
