Nvidia, Microsoft, Meta and Other Tech Giants Urge US Administration to Support Open-Weight AI
A coalition of 25 leading technology companies, including Nvidia, Microsoft, Meta, IBM, and Andreessen Horowitz, has penned a joint letter urging the US administration to support open-weight artificial intelligence (AI) models. The move comes as Washington considers tighter controls on advanced AI technologies amid growing concerns over China’s rapid progress in AI.
Nvidia CEO Jensen Huang, who has been an outspoken proponent of open-weight AI, shared his thoughts alongside the letter: “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.” He emphasized that open models strengthen safety and cybersecurity, accelerate innovation, and promote sovereignty.
The coalition argues that open-weight AI is crucial for maintaining America’s technological leadership. Open models allow developers to download trained parameters of an AI model, run them on their own infrastructure, and customize the application without depending entirely on the original developer. This approach contrasts with closed AI systems, which typically limit how users can interact with or modify the model.
The letter highlights that training advanced AI models can cost billions of dollars, making it prohibitive for most startups, governments, and universities. Open-weight models address this challenge by enabling developers to build upon existing AI infrastructure rather than starting from scratch. This approach also allows companies to choose a model that best suits their needs, reducing costs while reserving the most powerful and expensive models for critical tasks.
The authors of the letter argue that open-weight AI promotes healthy competition among AI developers and across various industries such as cloud providers, chipmakers, software companies, and application developers. Increased competition leads to faster innovation, lower costs, and broader economic benefits. As AI becomes integral to everyday business operations in sectors like healthcare, manufacturing, education, and scientific research, these benefits are becoming increasingly important.
One of the key advantages of open-weight AI is that it provides businesses with greater control over their data. Companies can keep their own data, customize models for specialized tasks, and deploy them wherever they see fit without relying on a single vendor. This autonomy also allows organizations to continue improving the model over time, retaining the value created through their AI investments.
Behind this debate lies a commercial reality: open-weight AI lowers the cost of building AI products, enabling startups and smaller companies to develop applications without spending billions on training frontier models. Conversely, proprietary AI platforms like OpenAI, Anthropic, and Google have an incentive to keep their most advanced models exclusive, fueling a broader discussion about whether the future of AI should be dominated by closed systems or supported by both proprietary and open-weight models.
The letter acknowledges the risk that open-weight AI could fall into the hands of bad actors but argues that openness can strengthen security. For instance, Hugging Face faced a cybersecurity incident where an autonomous AI agent breached their platform during a security evaluation. Initially relying on hosted frontier models for forensic analysis, they were blocked by safety guardrails and instead used the open-weight Chinese AI model GLM-5.2 to investigate the breach.
The coalition urges the US government to expand access to computing resources, invest in shared datasets and evaluation tools, and avoid regulations that may hinder the development or use of open models. They also warn against broad restrictions on AI training techniques such as distillation, arguing that legitimate improvement methods should not be conflated with intellectual property theft.
The debate over open versus closed AI is becoming one of the defining questions of the AI era. While closed models offer tighter control, open-weight models aim to spread AI capabilities across business domains, researchers, and developers. The coalition argues that this approach will help the US maintain technological leadership by encouraging innovation, increasing competition, and ensuring that AI benefits more sectors of the economy rather than a few companies.
This outcome could ultimately shape not only the future of AI in the US but also who leads the next phase of global AI development.
Thinking about building an AI product?
Get in Touch