The U.S. government wanted to protect national security by slowing down American AI models. Instead, it gave Chinese rivals a window to close the performance gap. Now corporate America is choosing cheaper Chinese models over domestic ones.
The background to this shift lies in a series of export controls imposed by the Trump administration, which were originally designed to block advanced chips and AI technology from reaching China. The unintended consequence was that Chinese labs, such as Zhipu, launched models that rival frontier labs while American firms faced delays and regulatory hurdles.
An export control directive from the Trump administration forced Anthropic to halt its Mythos 5 model for two weeks. OpenAI also restricted the rollout of its GPT 5.6 models at the government's request. "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," said venture capitalist Marc Andreessen.
Get your free investing masterclass bonus when you join Market Briefs, our free daily newsletter
The situation was described as "a pretty good wake-up call" by Sam Bresnick, a Georgetown researcher who studies security and emerging technology. David Sacks, former Trump crypto and AI czar, wrote: "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."
Zhipu's GLM 5.2 is an open-weight model, meaning its code is public and can be used freely. A token is a small unit of text the model processes.
Brian Armstrong, CEO of Coinbase, posted on X that his firm now uses open-weight models including GLM 5.2 and Kimi 2.7, and as a result has reduced its AI expenses by nearly half even while processing more tokens. Both Shopify and Airbnb have publicly praised Alibaba's Qwen 3 for helping them scale their AI capabilities. But the shift worries cybersecurity experts. "With the open-weight models, it's kind of the Wild West," Armadin co-founder Travis Lanham commented. Hed Kovetz, CEO of Silverfort, warned Chinese models could automate cyberattacks within months, adding: "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."
In response to Elon Musk's forecast that GLM 5.2 would achieve Fable-level performance by early next year, Zhipu founder Jie Tang said it "won't take that long."
U.S. policy makers are now debating whether to allow continued American access to Chinese AI models, given national security concerns about cybersecurity and hardware export controls. Companies will keep chasing the lowest cost, and the gap between U.S. and Chinese models may shrink further.
Subscribe to Market Briefs, our free daily newsletter, and claim your bonus investing masterclass