More than 25 technology companies, including Microsoft, Nvidia, Meta, IBM, Dell, Palantir and Hugging Face, have urged US policymakers not to impose broad restrictions on open-weight AI models, arguing that open AI ecosystems are essential for innovation and competitiveness.
The appeal comes as governments increase scrutiny of AI regulation and as Chinese companies release increasingly capable open-weight models.
Open-weight AI models allow developers to download and fine-tune a model's trained parameters on their own infrastructure instead of relying solely on cloud-based APIs. While they do not include the full source code or training data, they offer greater flexibility than proprietary AI systems.
The discussion has intensified following the release of Chinese models such as Moonshot AI's Kimi K3 and DeepSeek R1, which have performed strongly on industry benchmarks. Analysts say freely available open-weight models could accelerate AI adoption and influence which technologies businesses choose to build on.
The debate reflects two competing approaches to AI development: proprietary models controlled by individual companies and open-weight models that allow broader access for developers. The outcome of ongoing policy discussions could shape AI innovation, regulation and global competition in the years ahead.
Industry split
Companies such as Nvidia and Meta argue that open-weight AI promotes innovation and strengthens the AI ecosystem. Nvidia benefits from greater demand for AI hardware, while Meta has made open-weight models central to its AI strategy through its Llama family.
OpenAI and Anthropic, however, did not sign the letter, reflecting a different business model centred on proprietary AI services delivered through paid APIs.