Chinese company Moonshot AI released a new version of its Kimi model this week, perhaps sparking a wave of discourse about China and open source AI.
Moonshot said that while the Kimi K3 “still trails the most powerful proprietary models, Claude Fable 5 and GPT 5.6 Sol,” the new open source model “has demonstrated front-line performance across our evaluation suite and has consistently outperformed other models tested.” Independent analysis by Arena.ai and Vals AI shows that Kimi can compete with the flagship Frontier model.
The announcement, which coincided with Chinese President Xi Jinping’s speech at the World AI Conference in Shanghai, appeared to surprise Wall Street, sending the Nasdaq down about 1% on Friday as investors sold shares of chip companies such as Nvidia.
Many of the resulting posts from tech industry figures will sound familiar to those who remember the controversy after another Chinese company, DeepSeek, released its open source R1 model in January 2025. Except now, with the Trump administration’s tariff war with China, repeated fights over the national security threat supposedly posed by Anthropic, and a major AI company finally preparing to go public, everything seems to have escalated.
For example, David Sacks, the Trump administration’s former AI czar and current co-chair of the President’s Council of Advisors on Science and Technology, contrasted Kimmy’s progress with “an America in chaos. Politicians and bureaucrats are banning new data centers, tightening state regulations, and pushing new federal agencies to pre-approve cutting-edge models. This is how you lose the AI race.” (The news also gave him an excuse to investigate Anthropic and called Claude an example of the “woke lobotomist model” who is “the enemy of American competitiveness.”)
And former Uber CEO Travis Kalanick echoed complaints that the Chinese are “distilling” American AI models (i.e., being trained on the results).
“If distillation is not enforced, everyone must be able to distill from everyone else, or the American model will have one arm tied behind its back,” Kalanick wrote. (Of course, the American model is also based on the Chinese model, especially Kimi.)
Meanwhile, Dean Ball, head of strategic futures at OpenAI, said Kimi is a “very good model” whose performance “can’t be explained by distillation or anything like that.” He added, “Personally, I am surprised that the Chinese state continues to allow open source of such a good model considering the potential risks.”
In fact, Ball suggested that “the likely outcome of a world dominated by open models is full-blown AI communism.” Here, AI is “treated as a ‘public good’ that will ultimately be provided by the state as a kind of ‘digital public infrastructure’.”
“This future seems like a dystopian hellscape to me, but I’ve never met an advocate of an open model who doesn’t ultimately acknowledge that this is where it will all end,” Ball said. He even suggested that the Trump administration he served in would eventually realize that it would have to “create a large amount of regulatory risk associated with using the open China model.”
“There’s no need to ‘ban open source’ (one of the dumb topics in AI policy discussions),” Ball said. “Just tell all agencies to enact soft laws that create FUD (fear, uncertainty and doubt): ‘Federal Reserve Advisory Bulletin Finds Chinese AI Model May Have Backdoors.’ It doesn’t need much justification. “It creates enough regulatory risk for every regulated company to back down.”
But Shakeel Hashim, editor of Transformer, an AI-focused publication, argued that much of the concern is overblown because Kimi “is unlikely to have dangerous cyber capabilities” and if the Chinese government develops such capabilities, it would face “very similar incentives” that limit the open China model.
If you purchase through links in our articles, we may receive a small commission. This does not affect our editorial independence.
