A new artificial intelligence model from China, Kimi K3 from Beijing-based startup Moonshot, has surprised the U.S. tech industry. The model’s price is the highest yet for a Chinese AI offering, but it remains half as expensive as OpenAI’s high-performing GPT-5.6 Sol model, according to a Friday report by Bank of America research analysts. This market entry intensifies the rivalry between the world’s two biggest economies, directly challenging the revenue streams of established American AI corporations.
The Kimi K3 model topped Arena’s ranking for “front-end coding capability,” a key measure of AI performance. Anastasios Angelopoulos, co-founder and CEO of Arena, called it “the single biggest release of the year,” noting that open-source Chinese models are now surpassing closed U.S. models. More results continue to show K3 at the top of the pack, he said on social media.
The Capitalist Race for AI Dominance
This unveiling came just before Chinese President Xi Jinping’s opening address to the nation’s annual World Artificial Intelligence Conference in Shanghai. Xi spoke of AI development as a “symphony of global cooperation,” a sentiment that clashes with the reality of American-led restrictions. These restrictions have blocked China from accessing some of the world’s most advanced technologies, compelling China to build its own know-how and fueling the economic competition.
K3 follows another major AI model release last month from Chinese startup Zhipu, or Z.ai. Zhipu’s new flagship GLM-5.2 model is already widely used by software developers globally, performing nearly as well as top U.S. models at a lower price point. The market-shaking panic around K3 echoed the response to Chinese startup DeepSeek’s model release about 1 year ago, though tech analyst Patrick Moorhead called the K3 reaction an “overreaction shockingly similar” to DeepSeek’s. He acknowledged it could benefit parts of the broader AI industry but poses a direct revenue challenge to Anthropic and OpenAI.
During the conference, which runs until Monday, tech giant Huawei has also showcased its new Atlas 950 SuperPoD AI computing system. This signals China’s increasing ability to amass domestic hardware, despite U.S. import restrictions on chipmakers like Nvidia. Moonshot, the developer of K3, is a partner with Huawei, though it hasn’t disclosed the hardware used for K3.
Accusations of Surplus Extraction
U.S. politicians and major U.S. AI companies, including Anthropic and OpenAI, have accused Chinese AI models of “illicit distillation” to extract their technologies. Beijing has labeled these claims “groundless.” In February, Anthropic specifically accused DeepSeek, Moonshot, and MiniMax, another China-based AI lab, of engaging in campaigns to “illicitly extract Claude’s capabilities to improve their own models.”
Anthropic detailed that distillation involves training a less capable model on the outputs of a stronger one. While legitimate in some contexts, it becomes problematic when competitors “use it to acquire powerful capabilities from other labs in a fraction of the time, and at a fraction of the cost, that it would take to develop them independently.” This practice highlights the relentless pursuit of cost reduction and accelerated profit generation within the tech sector.
The process isn't one-sided. San Francisco-based startup Anysphere, creator of the popular coding tool Cursor, has admitted that one of its top products was based on Moonshot’s K2.5 model. Elon Musk’s SpaceX is planning to finalize a deal to purchase Cursor for $60 billion later this year, demonstrating capital consolidation around these rapidly developed technologies.
Moonshot co-founder and CEO Yang Zhilin earned his Ph.D. 7 years ago at Carnegie Mellon University, where he made fundamental contributions to machine learning. His former adviser, Russ Salakhutdinov, a former director of AI research at Apple, celebrated K3 as a “huge win for the open-source community.” Developers who build “open-source” AI make key components accessible for anyone to examine, modify, and build upon. Proponents argue this promotes innovation, while critics warn that publicly accessible powerful AI models pose safety and security dangers, often a pretext for maintaining proprietary control.