Meta CEO Mark Zuckerberg has voiced opposition to blocking Chinese AI models in the US, arguing it would not be an effective solution in the technological race. This was reported by the Financial Times.
According to him, banning advanced models developed in China would not be effective. He believes that American companies should systematically identify their own weaknesses and barriers to better compete with Chinese developers.
Debate Intensifies Over Kimi K3
The statement comes amid discussions in Washington about potential restrictions on Chinese models with open weights. According to Reuters, the debate intensified following the release of Kimi K3 by Beijing-based Moonshot AI.
Developers have introduced it as the world’s first open model of the 3T class with 2.8 trillion parameters, native vision, and a context of 1 million tokens. According to the project’s own assessments, it ranked behind only proprietary models Claude Fable 5 and GPT 5.6 Sol.
Some US officials and companies believe that Chinese labs may have used distillation. In July, Michael Kratsios, Director of the White House Office of Science and Technology Policy, accused Moonshot AI of this. That same month, US Treasury Secretary Scott Bessent stated that authorities would examine Chinese open-source models for signs of intellectual property theft.
He later clarified that in cases of “covert industrial distillation attacks,” sanctions and inclusion in export restriction lists are possible.
Industry Defends Open Weights
Zuckerberg’s stance aligns with a broader campaign to defend models with open weights. The Linux Foundation announced that it joined a letter alongside Cisco, Dell Technologies, IBM, Meta, Microsoft, Nvidia, and other companies.
As AI integrates across the global economy, how do we ensure it’s safe, transparent, and trustworthy?
At the Linux Foundation, we believe the answer is openness. Open-weight models give researchers & enterprises the visibility needed to inspect, audit, and fix vulnerabilities… pic.twitter.com/KD6Gr0sfdm
— The Linux Foundation (@linuxfoundation) July 27, 2026
“As AI integrates into the global economy, how can we ensure its safety, transparency, and reliability? At the Linux Foundation, we believe the answer lies in openness. Open-weight models provide researchers and enterprises with the visibility needed to inspect, audit, and address vulnerabilities before they can be exploited by malicious actors,” the post states.
Nvidia CEO Jensen Huang also spoke against bans. In an interview with Axios, he stated that good Chinese open models should be utilized, and the campaign to block them could harm the US itself.
Meanwhile, the White House is developing a mechanism for pre-screening the most powerful AI systems. In a June directive, authorities tasked the development of a voluntary framework allowing developers to provide the government with access to frontier models up to 30 days before release.
In July, participants in a march in San Francisco demanded that OpenAI, Anthropic, and Google DeepMind coordinate to halt the training of more powerful models. They suggested redirecting the freed resources towards safety.
That same month, former OpenAI researcher Daniel Kokotajlo proposed slowing the AI race until 2040.
