Moonshot's Kimi K3 Pauses Signups After 48h Demand Surge Maxes GPU Capacity

Moonshot AI paused new subscriptions to its Kimi K3 model just days after launch, as surging demand strained its computing capacity amid heightened competition between Chinese and US AI developers. Moonshot launched Kimi K3 last week as a 2.8 trillion-parameter model, making it the largest open-weight system released to date. Over the past 48 hours, demand has pushed close to the limits of Moonshot's current capacity, and the company is temporarily pausing new subscriptions to protect the experience of existing subscribers.

Kimi K3's 2.8 trillion parameters make it the world's largest open-source AI model, and the model aims to challenge more advanced AI models. Benchmark tests showed Kimi K3 outperforming Anthropic's Opus 4.8 model and OpenAI's ChatGPT 5.5 on most coding tasks, though it is still less powerful than Anthropic's cutting-edge Fable model. Moonshot underestimated how popular K3 would become, and the model is unusually demanding in terms of compute, which makes allocating capacity both difficult and costly.

The launch coincides with similar moves by other Chinese AI firms, including Z.ai and MiniMax, which are releasing increasingly capable models at lower costs, with Alibaba's recent debut of Qwen3.8-Max-Preview, a 2.4-trillion-parameter model. For infrastructure planners and cloud operators, Kimi K3's subscription freeze within 48 hours of launch signals that GPU capacity remains the hard constraint even for well-funded Chinese labs. Full model weights for Kimi K3 are expected to be released as open-source by July 27, which may help alleviate demand pressures on Moonshot's infrastructure. The visible pattern: cutting-edge model quality no longer gates adoption—GPU availability does.