How does Kimi compare to U.S. AI models in capability?

Kimi appears to rival the intelligence of models from OpenAI and Anthropic. It is described as nearly as good as an Anthropic model that the U.S. government deemed powerful enough to briefly shut down over national security concerns.

Why are Trump advisors so divided over Kimi?

David Sacks believes open AI is better and the government shouldn't eliminate open-source competition; Emil Michael and others argue national security demands government control over AI. The split reflects disagreement over whether to let market forces work or intervene to protect U.S. companies and national interests.

What has the U.S. government done to address Chinese AI companies copying American models?

In April, the Trump administration announced a series of efforts to curb distillation—the practice of training AI models on the outputs of existing models. OpenAI and Anthropic had long requested government help to stop this practice.

Large Language ModelsAI Regulation & PolicyAI Business & IndustryMIT Technology Review AIJul 20, 2026Jul 21, 2026, 04:00 JST

China's free AI model Kimi divides Trump's advisors on strategy

Send on LINE

3 Key Points

What happened
Moonshot, a Chinese AI company, released Kimi last week—a free, open-source model that appears to match the capability of paid models from OpenAI and Anthropic. The release sparked public disputes among Trump administration officials and advisors over how to respond, with David Sacks (the president's former AI czar) calling U.S. companies' models "lobotomized" and "woke," while Emil Michael, a Pentagon official, attacked OpenAI advisor Dean Ball as a "supreme village idiot."
Why it matters
Kimi's existence undermines the business case for paying U.S. AI companies, threatening a sector that is driving an outsized share of economic growth and propping up stock valuations. The model exposes a fundamental split in Trump's circle: some, like Sacks, believe open AI is better and the government shouldn't protect corporate interests; others, like Michael, argue national security demands government control over AI release and use. This division leaves Trump without a clear policy response to Chinese models that come with fewer usage restrictions and appeal to users seeking free alternatives.
What to watch
The White House has launched a new review process to vet AI models' security before release—which Ball criticized as a "de facto licensing regime for frontier AI." The unresolved question is how China trained Kimi despite limited computing power; Moonshot may have used distillation (training on outputs from existing U.S. models), a practice the Trump administration said in April it would work to curb. Kimi is described as nearly as capable as an Anthropic model that the U.S. government briefly shut down over national security concerns.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, in your inbox every morning.

Get news like this every morning, free

Context & Analysis

The release of Kimi reflects the erosion of U.S. chip export controls that had been a cornerstone of Biden-era China policy. Trump made the controversial decision to allow Nvidia to sell more chips to China in exchange for government revenue, loosening restrictions that were meant to keep advanced AI capabilities out of Chinese hands. Yet Kimi's emergence suggests that even with limited computing power, Chinese companies are finding ways to build competitive models—possibly through distillation, where they train on outputs from U.S. models like those from OpenAI and Anthropic.

The dispute among Trump's inner circle reflects deeper philosophical and political tensions. Sacks, whose formal role ended in March, represents a libertarian strain that opposes corporate protectionism and favors fewer restrictions on AI. Michael and others closer to the Pentagon take a national-security-first stance, arguing that AI's power demands government vetting and control. The White House review process Ball criticized represents this latter position—a shift from Sacks's preference for open markets toward state gatekeeping. Neither approach has momentum to become settled policy, leaving the administration without a coherent response as Chinese models undercut U.S. companies' market position and threaten the sector's outsized contribution to economic growth.

FAQ

How does Kimi compare to U.S. AI models in capability?: Kimi appears to rival the intelligence of models from OpenAI and Anthropic. It is described as nearly as good as an Anthropic model that the U.S. government deemed powerful enough to briefly shut down over national security concerns.
Why are Trump advisors so divided over Kimi?: David Sacks believes open AI is better and the government shouldn't eliminate open-source competition; Emil Michael and others argue national security demands government control over AI. The split reflects disagreement over whether to let market forces work or intervene to protect U.S. companies and national interests.
What has the U.S. government done to address Chinese AI companies copying American models?: In April, the Trump administration announced a series of efforts to curb distillation—the practice of training AI models on the outputs of existing models. OpenAI and Anthropic had long requested government help to stop this practice.

MIT Technology Review AIRead Original Article

Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Get it free

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.