A coalition of leading technology companies, including Microsoft, Nvidia, Meta, and more than 150 startups, investors, researchers and industry organisations, has urged the US government to support the development of open-weight artificial intelligence (AI) models, warning that excessive restrictions could weaken innovation, hurt businesses and allow global competitors to gain an edge.
The appeal comes as policymakers in Washington debate new rules for advanced AI systems amid growing concerns over national security, cybersecurity and the rapid rise of Chinese AI companies. While governments are looking to regulate powerful AI technologies, the technology industry argues that open-weight AI has become a critical driver of innovation and economic growth.
In a joint letter to the US administration, the coalition said open-weight AI models are essential for maintaining America’s leadership in artificial intelligence. The signatories warned that limiting access to these models could have unintended consequences for thousands of businesses that rely on them to build AI-powered products and services.
Unlike fully closed AI systems, open-weight AI models make their trained parameters, or “weights”, available to developers under specific licensing terms. This allows organisations to study, fine-tune and customise models for different applications without building an AI system entirely from scratch. Although they are not completely open-source, open-weight models offer significantly greater flexibility than proprietary AI platforms.
The companies argue that this accessibility has fuelled rapid progress in generative AI, enabling startups, universities, enterprises and independent developers to create innovative tools at a much lower cost. Instead of investing billions of dollars to train large language models, organisations can adapt existing open-weight models for industries such as healthcare, education, finance, manufacturing and scientific research.
The debate has intensified following the rapid growth of Chinese AI companies, many of which have released competitive open-weight models that are attracting developers worldwide. Industry leaders believe the United States must continue encouraging open AI innovation if it wants to remain competitive in the global AI race.
According to the coalition, restricting access to open-weight AI models while rival nations continue advancing their own technologies would place American companies at a disadvantage. They warned that hundreds of businesses built around open AI models could face serious challenges if new regulations limit access to these technologies.
The letter emphasises that open-weight AI has become the foundation of a growing ecosystem of startups and developers. These companies use accessible AI models to create applications ranging from customer service assistants and software development tools to medical diagnostics, multilingual communication platforms and scientific research systems.
Nvidia, one of the world’s leading AI chip manufacturers, has strongly supported the open-weight approach, saying wider access to AI models encourages innovation across the technology industry. The company believes developers can build better products when they have the freedom to experiment, improve existing models and create specialised AI solutions for different sectors.
Microsoft has also highlighted the importance of open-weight AI as part of responsible AI development. The company says giving developers access to model weights promotes transparency, collaboration and innovation while allowing businesses to adapt AI systems to local languages, regulations and customer needs.
According to Microsoft, open-weight models strike a balance between accessibility and responsible deployment. While developers can inspect and fine-tune the models, licensing terms continue to govern how they are distributed and used, helping address concerns around misuse.
Meta has been one of the strongest advocates of open-weight AI through its Llama family of large language models. The company has repeatedly argued that making advanced AI more widely available prevents innovation from being concentrated in the hands of only a few large technology companies.
Supporters of open-weight AI also believe the approach improves security. By allowing researchers and developers to examine AI models, vulnerabilities can be identified more quickly, leading to stronger safeguards and safer deployment. They argue that broader scrutiny often results in faster improvements than closed systems, where only the original developers have complete visibility into the technology.
However, governments remain cautious about the rapid advancement of AI. Policymakers are concerned that powerful AI systems could be misused for cyberattacks, misinformation campaigns, fraud or the creation of harmful content. These concerns have prompted discussions on introducing stricter regulations for advanced AI models.
