A new artificial intelligence model from China is rattling US markets. Chinese firm Moonshot unveiled Kimi K3, a model the New York Post reports has capabilities close to those of American leaders Anthropic and OpenAI.
The reaction on the markets was swift. According to MSN, stocks in Asia and the United States fell on Friday, with the Nasdaq dropping 1%, as the announcement intensified concerns that the heavy AI spending driving this year's rally may be overextended. Business Insider framed the model as a possible "next DeepSeek" — a reference to an earlier Chinese release that spooked investors — saying the surprise breakthrough is rattling US market heavyweights. Futurism reports that Kimi K3 shot to number one on the charts.
There are questions about how the model was built. Wccftech reports that in at least one conversation, Kimi K3 identified itself as Anthropic's Claude — which the outlet describes as betraying "distilled" origins, meaning it may have been trained partly on the outputs of existing US models.
Not everyone sees the news as purely negative. Investor Gavin Baker argued on Techmeme that models like Kimi K3, alongside Grok 4.5 and Muse 1.1, may prevent a handful of dominant frontier labs — which he says enjoy inference margins around 90% — from squeezing other layers of the AI ecosystem. Baker called Kimi K3 a potentially important inflection point, though he noted it could be negative for Anthropic and OpenAI specifically.
Why it matters: if a Chinese lab can match top US models at lower cost, it challenges the assumption that a few American companies will control AI's most valuable technology — and the trillions in market value built on that bet.