Moonshot AI's Kimi K3 drives $50 billion valuation and fuels US‑China AI competition

Moonshot AI, a Beijing‑based startup, unveiled its open‑weight language model Kimi K3 in July 2026. The model contains 2.8 trillion parameters, making it the largest open‑source model worldwide, and benchmark tests showed it beating or matching leading US models such as Anthropic’s Claude Fable 5 and OpenAI’s GPT‑5.5 in coding and reasoning tasks.

Following the launch, demand surged, prompting Moonshot to temporarily suspend new user subscriptions on 19 July because its GPU capacity could not keep up with the influx. The company reported annual recurring revenue of $300 million in June, up from $200 million in April, and plans a final pre‑IPO funding round valued at $31.5 billion, with a target valuation of up to $50 billion before a possible Hong Kong listing as early as 2026.

The rapid rise of Kimi K3 has triggered geopolitical concern in the United States. Officials are debating adding Chinese AI labs to the Commerce Department’s Entity List and considering broader restrictions on Chinese AI models, while analysts describe the development as a “Sputnik moment” for U.S. policy. Investors such as Alibaba and Tencent have stakes in Moonshot, highlighting the broader Chinese tech ecosystem’s push into AI.

Chinese social media dubbed founder Yang Zhilin a “millennial genius,” with a Weibo hashtag reading “Post‑90s Tsinghua entrepreneur shakes Wall Street.” The model’s low per‑token cost—around $0.95 compared with $2.75 for comparable U.S. models—adds pressure on American providers.

Overall, Moonshot’s funding push, the Kimi K3 performance, and the ensuing U.S. policy response mark a notable escalation in the global AI race.