Open-source AI models from China are once again causing a crisis in America, leading to accusations of stolen training data and threats of closing the border to foreign-origin tech. With tensions high, some tech executives are looking to broker peace before the Trump administration declares war on open-source on behalf of America’s precious closed-source models produced by labs like OpenAI and Anthropic.
Nvidia CEO Jensen Huang went so far as to create an X account to post about the importance of open source, which is a real display of bravery because you could not come up with a worse platform to discuss the nuances of this argument.
“AI will transform every industry, power every company, and be built by every country,” Huang wrote in his very first post. “Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty. The world needs both frontier closed models and frontier open models.”
Nvidia’s chief executive also shared a letter titled “Open Weights and American AI Leadership,” which made the case to policymakers not to clamp down on open-source models—instead, urging them to make sure the frontier model race stays open by “avoiding premature restrictions on open models that stifle competition or drive innovation overseas.”
The letter argued that “Open weights also strengthen competition and competition is what keeps the gains of AI broadly shared rather than concentrated in a few hands,” warning that “Relying solely on closed models is not inherently safe: they can be breached, misused, or fail in ways that outsiders cannot detect.”
Joining Huang and Nvidia in signing the letter were 24 other companies, including some of the usual suspects when it comes to pushing for open source—Mozilla, Hugging Face, the Linux Foundation—and some that seem a bit surprising at first blush, like Meta, Microsoft, and Palantir.
The same force explains both their presence and OpenAI and Anthropic’s absence: self-interest. Palantir is model-agnostic, so while its founder is certainly a China hawk, he’s also self-interested. Meta and Microsoft are trying to play catch-up in the AI race, and Microsoft has reportedly been mulling over the possibility of embracing DeepSeek and other open-source models to keep their Copilot costs down while they build out their own model. Meta, meanwhile, has repeatedly promised it would open source its models—though Muse Spark, its current flagship model, is still closed.
Nvidia’s position is not without its self-interest, either. The company doesn’t really care whose model is running on its chips, as long as everyone is reliant on them. It stands to be one of the companies that gets hit should the Trump administration crack down on Chinese AI firms, given that one of the preferred tools thus far has been restricting access to American-made CPUs and GPUs, which would mean less business for Nvidia.
Anthropic and OpenAI (note the “open” in its name), meanwhile, are sounding the alarm bells over these open models, doing their best to cause a stir by playing up the security risks and leaning on the China-phobia that so many in the Trump administration are afflicted by. It seems like it’s working, given that Trump is reportedly considering new restrictions on Chinese models.
If Huang’s letter doesn’t manage to get through to lawmakers, at least he’s now got a way to post through it on X, the everything app.