U.S. AI Crackdown Risks Opening Door for Chinese Competitors

The Trump administration’s restrictions on leading U.S. artificial intelligence models could give Chinese AI developers an opportunity to close the gap with American companies such as Anthropic and OpenAI, as China’s technology firms release increasingly competitive models.

Anthropic was allowed by the White House on Friday to release its Mythos 5 model to select companies and federal agencies after a two-week pause linked to an export control directive, although its Fable 5 model remains unavailable. OpenAI also said it would limit the rollout of its GPT 5.6 models following a government request.

The developments come as the United States seeks to balance rapid AI development with national security concerns. U.S. officials and technology executives have argued that excessive restrictions could slow domestic innovation and benefit China, which has been rapidly advancing its AI capabilities.

Chinese companies are now releasing models that researchers say are approaching the performance of leading U.S. systems in some areas. Zhipu’s open-weight GLM 5.2 model, launched earlier this month, has reportedly matched top U.S. AI labs on some cybersecurity benchmarks, including performance comparable to Anthropic’s Mythos model.

“Many smart people/AI insiders are saying GLM-5.2 is the first Chinese AI model to match and often beat the American big lab public AI models with no compromises,” venture capitalist Marc Andreessen wrote on X. “Incredible timing given current events.”

Sam Bresnick, a research fellow at Georgetown’s Center for Security and Emerging Technology, described the situation as “a pretty good wake-up call.”

Jefferies strategist Christopher Wood, citing industry sources, said in a report to clients that GLM 5.2 “is almost equal to Anthropic as a competitor for the corporate market and is just one quarter of the cost in terms of cost per token.”

Former Trump crypto and AI adviser David Sacks, who has criticized Anthropic’s approach to AI safety, also weighed in on the debate, writing: “A year ago, President Trump declared that America was in a global AI race and that the way to win it was to be pro-innovation, pro-infrastructure, pro-energy, and pro-export. President Trump was exactly right; we deviate from that strategy at our peril.”

Representatives from Anthropic, OpenAI and the White House did not respond to requests for comment.

The shift comes as companies move away from unrestricted AI spending toward greater focus on efficiency and cost. That trend could benefit lower-cost Chinese models, particularly open-weight systems that companies can run on their own infrastructure.

Earlier this month, Flo Crivello, CEO of AI startup Lindy, said the company moved all of its traffic from Anthropic’s Claude models to Chinese AI company DeepSeek’s cheaper open-weight alternatives.

“We did it, and you could see that cost curve go down, like, crash to the ground,” Crivello told CNBC.

Chinese open-weight models are becoming increasingly accessible to businesses globally because developers can download and operate them without relying entirely on external cloud providers.

“With the open-weight models, it’s kind of the Wild West,” said Travis Lanham, co-founder of AI security startup Armadin, which is testing GLM 5.2 and Moonshot AI’s Kimi K2.7 model.

Lanham said the systems have shown improved capabilities for cybersecurity applications, including analyzing reconnaissance data and generating exploit code.

The rise of Chinese AI models has intensified debate over U.S. technology controls aimed at limiting China’s access to advanced computing capabilities. Washington has restricted exports of advanced AI chips from companies including Nvidia and Advanced Micro Devices, while also imposing restrictions on certain Chinese technology firms over national security concerns.

Major companies are also increasingly experimenting with Chinese AI models. Shopify and Airbnb have highlighted Alibaba’s Qwen 3 for scaling AI features, while Coinbase CEO Brian Armstrong said his company uses open-weight models including GLM 5.2 and Kimi K2.7, cutting AI spending while increasing usage.

Cybersecurity remains a major concern among industry experts, who warn that increasingly capable open models could automate parts of cyberattacks.

“If the U.S. government does not let the industry take advantage of this opportunity to get ready, then when the Chinese models reach a similar level, no one will be prepared,” said Hed Kovetz, CEO of cybersecurity startup Silverfort.

Stay ahead of the stories shaping our world. Subscribe to Impact Newswire for timely, curated insights on global tech, business, and innovation all in one place.

Dive deeper into the future with the Cause Effect 4.0 Podcast, where we explore the ideas, trends, and technologies driving the global AI conversation.

Got a story to share? Pitch it to us at info@impactnews-wire.com and reach the right audience worldwide