Jensen Huang, CEO of US chip giant Nvidia, posted on social platform X for the first time to voice strong support for open-source artificial intelligence (AI) models, emphasizing that the world needs both cutting-edge closed-source models and cutting-edge open-source models.
On Friday, July 24 (local time), Huang wrote on X: “AI will transform every industry, empower every company, and be built by every nation. Open-source models can strengthen security and cybersecurity, accelerate innovation and adoption, and enable sovereignty. The world needs both cutting-edge closed-source models and cutting-edge open-source models.”
In his tweet, Huang attached a joint open letter titled 'Open Weights & U.S. Leadership in AI,' jointly signed by more than 20 tech giants and renowned institutions including Nvidia, Meta, Microsoft, Palantir, IBM, Hugging Face, and others.
Notably absent from the letter were the leading U.S. closed-source AI labs OpenAI, Anthropic, and Google.
The open letter urges U.S. policymakers to avoid imposing “premature restrictions” on open-weight AI models, warning that such actions could stifle competition or drive innovation overseas. The letter emphasizes that open-weight models allow users to download, modify, and run the models on their own infrastructure, thereby enhancing market competition and ensuring the benefits of technology “are widely shared, not concentrated in the hands of a few.”
The letter acknowledges that open-source models do carry risks—once released, models move beyond the control of their developers—but argues that prohibition is not the right approach. Defenders must also have access to models with similar capabilities to detect, simulate, and respond to new types of threats.
The letter adds, “Relying solely on closed-source models is inherently unsafe: they can also be hacked, abused, or fail in ways that outsiders cannot detect. Concentrating advanced AI capabilities behind a handful of closed-source models actually exacerbates this risk.”
This open letter comes at a time when there is an intense debate within Silicon Valley and Washington over open-source versus closed-source ecosystems. Currently, China's open-source large AI models are rapidly rising, posing a direct challenge to top U.S. companies like OpenAI and Anthropic, which primarily develop closed-source models.
Previously, U.S. officials accused the K3 model of distilling Western models, and are considering imposing sanctions and restrictions on Chinese models.
However, the joint letter cautions policymakers against “blanket restrictions” on distillation technology. The letter states that distillation technology—which trains new models using the outputs of existing models—has a “long-standing tradition” in AI improvement and validation, and legitimate model development techniques should not be conflated with illegal infringement.
The letter concludes: “Our leadership in the field of artificial intelligence should not be judged solely on individual cutting-edge AI models, but on whether America can build a strong, open ecosystem capable of permeating every industry. This is crucial for creating opportunities for innovation and prosperity across the United States.”