Microsoft's open-weight AI push is an Azure play as 25 companies urge US against restrictions

Microsoft, Nvidia, Meta, Mistral, and more than 20 other companies published an open letter on Friday urging U.S. policymakers not to impose broad restrictions on open-weight AI models. The letter, titled "Open Weights and American AI Leadership," argues that leadership will be judged by the strength of an open ecosystem that diffuses AI into every sector, not by a single frontier model.

The letter has 25 signatories, including Palantir, Perplexity, Hugging Face, Mozilla, the Linux Foundation, Dell Technologies, IBM, and Andreessen Horowitz. Notably absent are OpenAI, Anthropic, and Google, which have each pushed for tighter controls on open-weight models citing safety and national security concerns.

Microsoft's cost-saving calculus

For Microsoft, the open-weight push is a direct strategy to boost its Azure cloud business. The company makes money from cloud infrastructure and the lock-in created by its Windows and Office ecosystem. More models running on Azure mean fewer reasons for customers to use rival platforms. Microsoft is also replacing OpenAI and Anthropic models in GitHub Copilot, Excel, and Outlook with its in-house MAI family of models. Independent benchmarks put MAI well behind OpenAI and Anthropic, roughly matching DeepSeek V3.2, according to The Decoder. Microsoft claims MAI performs just as well or better, but only compares it against smaller, less powerful models like GPT-5.4 Mini and Anthropic's Haiku.

• Microsoft's MAI models run on older Nvidia GPUs like the H100 and A100 instead of requiring the latest accelerators, "significantly lowering the cost of deployment for Microsoft," the company said on its blog.
• The letter frames distillation, in which a smaller AI model learns from a more capable model's output, as a legitimate way to build AI, reflecting a tradition of building upon existing technologies.
• Chinese providers have faced particular criticism for using distillation, and the Trump administration reportedly plans to take action against Chinese open-weight models.

Implications for cloud competition and regulation

Microsoft CEO Satya Nadella wrote on LinkedIn that the key is using "the right model for each task" while keeping costs down. He had previously warned that a few AI models could absorb the value of entire industries. The company's shift to open-weight aligns with its interest in a fragmented model market where no single AI lab becomes powerful enough to threaten Microsoft's cloud and operating system businesses. The letter also serves as a preemptive strike against potential U.S. restrictions on Chinese AI models, which some in Washington have proposed in response to rapid advances from Chinese labs like DeepSeek.

What comes next: The letter is a signal to the Trump administration as it weighs response to Chinese AI. The signatories are betting that the U.S. will avoid a sweeping ban on open-weight models, favoring instead a more targeted approach. For Microsoft, the immediate payoff is clearer: lower costs, higher margins, and more Azure customers locked into its ecosystem. Whether those customers benefit from weaker AI at the same price remains an open question.