Dario Amodei, co-founder and CEO of Anthropic, at the AI Impact Summit in New Delhi, India, on Thursday, February 19, 2026.
Prakash Singh | Bloomberg | Getty Images
The Trump administration’s crackdown on Anthropic’s leading artificial intelligence models appears to be a gift to the country’s biggest opponent in the AI race: China
After a two-week shutdown due to an export control directive, Anthropic was allowed by the White House on Friday to release its powerful Myth 5 model to some companies and federal agencies, but the Fable 5 model remains off the market. OpenAI also said Friday that it would limit the rollout of its GPT 5.6 models at the government’s request.
The two leading US AI model developers are in a race against each other and against technology giants Google to develop the most advanced technology, with the US government opening the door to rapid AI development by limiting regulatory hurdles. Doing otherwise, according to many tech executives and Trump administration officials, would limit domestic AI to the benefit of China, which has quickly caught up with the United States
However, as Anthropic addresses the US government’s national security concerns, Chinese companies are launching models that can compete with Frontier Labs in some features. According to researchers, Zhipu’s GLM 5.2, released earlier this month, can compete with the best US labs in some cyber benchmarks and even rival Mythos’ capabilities.
“Many smart people/AI insiders say that GLM-5.2 is the first Chinese AI model to match and often uncompromisingly outperform the public AI models of major American labs,” wrote venture capitalist Marc Andreessen in a post on X over the weekend. “Incredible timing given current events.”
Sam Bresnick, a research fellow at Georgetown’s Center for Security and Emerging Technology, called the recent developments “a pretty good wake-up call,” and Jefferies strategist Christopher Wood, citing industry sources, wrote in a report to clients that GLM 5.2 “is almost on par with Anthropic as a competitor for the enterprise market, at just a quarter of the cost per token.”
Even former Trump crypto and AI czar David Sacks, who has been a vocal critic of Anthropic’s approach to AI security, penned a cryptic post
“A year ago, President Trump declared that America was in a global AI competition and that the way to win it was to champion innovation, infrastructure, energy and exports,” Sacks wrote. “President Trump was exactly right; we deviate from this strategy at our peril.”
Representatives for Anthropic, OpenAI and the White House did not respond to requests for comment.
The Chinese conundrum touches on big companies like America, which are transitioning from an era of so-called tokenmaxxing – allowing developers to spend without restrictions on AI – to a focus on efficiency and return on investment. This also plays into China’s hands.
Earlier this month, Flo Crivello, CEO of AI startup Lindy, transitioned his company away from Anthropic’s Claude models and shifted 100% of its traffic to DeepSeek, a Chinese company that makes cheaper, open-weight alternatives.
“We did it, and you could see the cost curve going down, like it was falling to the bottom,” Crivello told CNBC in a story published last week.
Table of Contents
“Wild West”
Chinese AI developers easily reach US users because of how easy it is for a company to download open-weight models and run them on their own servers without relying on a third-party cloud.
“It’s kind of the Wild West with the open-weight models,” said Travis Lanham, co-founder of AI security startup Armadin, which is experimenting with GLM 5.2 and another Chinese model, Moonshoot AI’s Kimi K2.7.
Lanham said the models demonstrate improved capabilities for cybersecurity use cases such as analyzing intelligence data and creating exploit code for customers.
Whether the US authorities will continue to allow this is increasingly questioned in political circles due to the way the two largest economies handle each other’s sensitive hardware. For years, the U.S. government has gone to great lengths to prevent cutting-edge AI innovations from falling into China’s hands, through U.S. export controls on AI chips Nvidia and advanced micro devices. The US also banned American companies from using Huawei equipment on national security grounds.
Last year, the US cleared Nvidia’s H200 chip – the same model used by US companies – for export to the China region. But Nvidia said earlier this year that it had not yet generated revenue from the chips and did not know whether China would allow imports of its products.
As for GLM 5.2, Tesla And SpaceX Founder Elon Musk wrote in a post
Zhipu founder Jie Tang wrote in a response to Musk: “It won’t take that long.”
It’s not just niche brands that use Chinese models. Big companies like it Shopify And Airbnb have touted the benefits of Alibabais Qwen 3 for scaling AI functions. Coinbase CEO Brian Armstrong wrote last week
Cybersecurity is the top concern for many industry professionals. Some open-weight models can already automate many phases of a cyberattack, and Hed Kovetz, CEO of industry startup Silverfort, fears they are just months away from carrying out a full operation.
“If the U.S. government does not allow the industry to take advantage of this opportunity to prepare, no one will be prepared when the Chinese models reach a similar level,” he said.
—CNBC’s Deirdre Bosa and Ashley Capoot contributed to this report.
REGARD: How Z.ai is approaching the American AI frontier
Choose CNBC as your preferred source on Google and never miss a moment from the most trusted name in business news.
https://www.cnbc.com/2026/06/30/white-house-ai-china-crackdown.html
