Meta CEO is not convinced if OpenAI & Anthropic really want Chinese AI models in US

Meta CEO Mark Zuckerberg has entered the intensifying Washington debate over whether Chinese AI models should be restricted in the US, wanting that banning them would not be ‘an effective solution’. In an interview with Financial Times, Zuckerberg said the US should compete by identifying bottlenecks and roadblocks rather than blocking rivals outright. He also cautioned against ‘regulatory capture’ by American frontier labs such as OpenAI and Anthropic, which he suggested could use policy to throttle domestic competition. Zuckerberg’s remarks appeared to be a broadside against OpenAI and Anthropic, which dominate the chatbot market with proprietary closed models. He said there were “mixed signals” about whether frontier labs truly want open-source models to succeed. “Are the frontier labs going to want an open-source model to succeed? I think that there have kind of been some mixed signals on that,” he told the FT.The increasing Chinese competition The comments from Zuckerberg comes after Beijing-based Moonsht AI released its Kimi K3 model, whose capabilities are nearing those of OpenAI and Anthropic. US officials have also accused Moonshot of covertly training its model using distillation from American rivals, with Treasury Secretary Scott Bessent warning that sanctions are on the table for labs that cross into intellectual property theft.Open vs closed modelsZuckerberg reiterated his support for “AI for everyone” and open models that users can customize and run on their own hardware — a popular approach among Chinese developers. He warned that AI risks becoming centralized in the hands of a few powerful groups. While Anthropic CEO Dario Amodei has called for stronger regulation and Google DeepMind’s Demis Hassabis has proposed a self-regulatory body, Zuckerberg said restrictions on powerful tools were not the right answer. He argued that open models could help patch cybersecurity vulnerabilities faster than closed ones.Zuckerberg’s intervention highlights a growing split among U.S. tech leaders over how to respond to Chinese competition and regulate AI. While rivals push for tighter controls, he is positioning Meta as a champion of open-source approaches — even as his company develops its own closed model, Muse Spark, citing safety concerns.Mark Zuckerberg makes an offer to Anthropic CEO Dario Amodei In related news, recently it was reported that Meta is in early talks to lease computing power to Anthropic in a deal that could be worth as much as $10 billion over two years, according to a New York Times report. If it closes, it would hand Mark Zuckerberg a brand-new line of business and pull Dario Amodei's AI lab deeper into a scramble for compute that shows no sign of easing.The move borrows straight from Elon Musk's playbook. Musk turned SpaceX's Colossus data centers into a rental business earlier this year, and Anthropic became its anchor tenant. Now Zuckerberg wants a slice of the same action, using the very data centers his investors keep grumbling about. Anthropic pitched the idea first, in JuneThe NYT reports that Anthropic proposed the deal in June, and Meta is weighing it. Anthropic would pay in monthly instalments across two years, with either side free to opt out early. The specifics are still moving around.Get the latest technology news and updates. Download the TOI App.