Nvidia CEO Jensen Huang Defends Open AI Models in First X Post
Key Takeaways
- •Jensen Huang published his first-ever X post to share a letter, signed by Nvidia and 24 other organizations, arguing that open-weight AI models are critical to US leadership in artificial intelligence.
- •US officials are currently considering restrictions on Chinese open-weight models including Kimi K3, a 2.8-trillion-parameter model released by Moonshot AI on July 16.
- •OpenAI and Anthropic did not sign the letter, reflecting an industry divide between companies that rely on proprietary model control and firms like Nvidia whose hardware sales benefit from broad AI adoption.
- •The signatories contend that open models improve safety by allowing widespread code inspection and distinguish the practice of distillation from intellectual property theft.
- •A prior episode involving China's DeepSeek model in January 2025 caused Nvidia to lose nearly $600 billion in market value in a single day, intensifying the policy debate over open AI model regulation.
Nvidia CEO Jensen Huang used his first-ever post on X, formerly Twitter, to defend open-weight artificial intelligence models and urge Washington not to restrict them.
The appeal concerns open-weight AI models, not OpenAI, the similarly named company. Open-weight models are systems whose weights can be downloaded and reused for free. For Nvidia, the stakes are both philosophical and commercial: the chipmaker's GPUs are the dominant hardware for training and running AI models, and broader adoption of open models that developers can download and customize expands the market for computing infrastructure.
Nvidia Backs Open Models
Huang shared a letter on Friday signed by Nvidia and 24 other organizations, including Microsoft, Meta, and Hugging Face. The letter argues that open models are important to US leadership in artificial intelligence.
For my first post, I’m sharing a letter @NVIDIA signed on why open models matter. AI will transform every industry, power every company, and be built by every country. Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty.… pic.twitter.com/t02bi51N4C — Jensen Huang (@JensenHuang) July 24, 2026
For my first post, I’m sharing a letter @NVIDIA signed on why open models matter. AI will transform every industry, power every company, and be built by every country. Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty.… pic.twitter.com/t02bi51N4C
The signatories compare open models to open-source software, which now supports much of the internet. They describe open models as a foundation for American AI competitiveness and innovation. Meta has already released its Llama family of open-weight models, which have been downloaded hundreds of millions of times, and Hugging Face hosts hundreds of thousands of open models on its platform.
The timing is notable. About a year earlier, the US government supported open models in its AI Action Plan, describing them as a strategic asset for the country.
US officials are now considering restrictions on Kimi K3 and other Chinese models. Kimi K3 was launched on July 16 by China's Moonshot AI. With 2.8 trillion parameters, it ranks among the strongest models globally, whether open or closed.
Concerns over Chinese AI models have surfaced before. In January 2025, a low-cost model from China's DeepSeek rattled Nvidia, and the chipmaker lost nearly $600 billion in market value in one day. At the time, that marked a record one-day loss. The DeepSeek episode underscored how quickly Chinese labs could close perceived gaps with US frontier models, adding urgency to the policy debate over whether openness helps or harms American interests.
Safety Arguments and Industry Split
The letter also presents a safety argument for openness. It says open models allow many teams to inspect code, identify flaws, and fix them quickly. By contrast, closed models remain controlled by a small number of firms, which the group describes as a weak point.
The signatories also distinguish distillation from theft. Distillation involves training one model using another model's output. The group characterizes that practice as normal research rather than stealing.
The issue remains contested in Washington. White House adviser Michael Kratsios has said Moonshot copied a US model to build Kimi K3.
Two major AI companies, OpenAI and Anthropic, did not sign the letter. Both companies have warned Washington that powerful Chinese open models pose risks. Their position reflects a fundamental fault line: companies whose business models depend on controlling proprietary frontier models have different incentives than those, like Nvidia, whose hardware sales benefit from the widest possible proliferation of AI development.
The AI sector is therefore divided over how open artificial intelligence should be. Chinese models, meanwhile, already outpace US competitors in daily use.
“Policymakers have an important opportunity to act… keeping the frontier plural by avoiding premature restrictions on open models that stifle competition or drive innovation overseas,” the signatories state in their letter.
