Zuckerberg opposes banning Chinese AI models: openness as a strategy, not a weakness
Meta CEO Mark Zuckerberg has taken a principled stance in the escalating tech war: he has strongly opposed blocking Chinese AI models in the United States. In his view, attempts to slow down competitors through administrative measures are a dead end. Instead of bans, American companies should focus on systematically analyzing their own weaknesses and removing internal barriers to compete on equal footing with Chinese developers.
The Spark of Conflict: Kimi K3 from Moonshot AI
The debate in Washington over potential restrictions on open-weight models gained new momentum following the release of the Kimi K3 model by Beijing-based Moonshot AI. The developers claimed to have created the world's first open model of the 3T class, featuring 2.8 trillion parameters, native vision, and a context window of 1 million tokens. According to their own benchmarks, Kimi K3 trails only proprietary giants — Claude Fable 5 and GPT 5.6 Sol.
This success sparked a wave of suspicion. Some U.S. officials and industry representatives believe Chinese labs may have resorted to distillation — a method where one model is trained on the outputs of another. White House Office of Science and Technology Policy Director Michael Kratsios has already publicly accused Moonshot AI of this, while Treasury Secretary Scott Bessent threatened sanctions and inclusion on export restriction lists if "hidden industrial distillation attacks" are uncovered.
Industry Unites Against Bans
Zuckerberg's position is not isolated. It is part of a powerful campaign in defense of open-weight models. The Linux Foundation, along with Cisco, Dell, IBM, Meta, Microsoft, Nvidia, and more than 20 other companies, signed an open letter. It emphasizes that openness is key to AI safety, transparency, and trust, allowing researchers and enterprises to inspect and audit systems before they can be exploited by malicious actors.
Nvidia CEO Jensen Huang also joined the chorus of voices against bans. In an interview, he stated outright that good Chinese open models should be used, and that the campaign to block them would backfire on the U.S. itself.
Washington's Double Game
Notably, amid these debates, the White House is simultaneously shaping control mechanisms. A June executive order tasks authorities with developing a voluntary framework allowing the government to access frontier models 30 days before their public release. This creates a curious precedent: on one hand, calls for openness; on the other, attempts to impose prior censorship.
Expert Opinion: The dispute over Kimi K3 is just the tip of the iceberg. Banning Chinese models is an attempt to plug a hole in the dam while water is already rushing in from all sides. Open source in AI is the new oil, and those trying to block it risk running out of fuel for their own technological growth. Instead of bans, American companies should ask themselves: why can't their own open models compete with Chinese ones?