Three Open Letters in Five Days: Microsoft, Anthropic and 1,324 Developers on Open-Weight Models

In late July 2026, the AI industry traded three open letters on the future of AI development. 235 companies led by Microsoft demanded protection for open models, Anthropic responded with its own position, and 1324 employees of leading AI labs urged the US to support an international slowdown of the race. All of it happened within five days.

What does the open-weights letter demand?

The letter "Open Weights and American AI Leadership," organized by Microsoft and published on July 24, 2026, was signed by 235 companies, including NVIDIA, Amazon, Y Combinator, The Linux Foundation and (joining later) OpenAI. It targets possible attempts by US authorities to ban or restrict open-weight models on "safety" grounds — a risk the industry began discussing after a series of restrictions on large AI systems.

• Publication date: July 24, 2026
• Number of signatories: 235 companies
• Signatories include: NVIDIA, Amazon, Y Combinator, Linux Foundation, OpenAI
• Central argument: closed models are just as vulnerable as open ones, and their monopolization is more dangerous

"Relying exclusively on closed models is inherently unsafe: they can be hacked, misused, or create situations that observers cannot detect.

Concentrating AI capabilities in the hands of a small number of providers creates single points of failure and weakens competition," the Microsoft-organized letter states.

An unexpected item was its support for distillation, the practice of training one model on another's outputs. The authors called it a "legitimate technique for improving models" with a long history in open-source software traditions, and warned regulators not to conflate distillation with copyright infringement.

Why did Anthropic refuse to sign?

Anthropic was not among the 235 signatories and on July 27, 2026 published its own position on open-weight models. CEO Dario Amodei hardened the company's rhetoric: he named authoritarian governments that could build "models more powerful than American ones," and the misuse of models for "cyberattacks or biological attacks," as the main threats.

Anthropic demanded a halt to "industrial-scale distillation operations" — a message directly opposed to the Microsoft letter. Amodei nonetheless added a caveat: "Anthropic has never called for a ban on open-weight models" — signaling that the company favors selective restrictions rather than a blanket ban.

Who called for slowing the AI race?

On July 28, 2026, the letter "Pacing the Frontier" appeared, signed by 1324 employees of leading AI companies. Among its authors are Jakub Pachocki (chief scientist at OpenAI), Ilya Sutskever (Safe Superintelligence Inc), Dario Amodei and Jack Clark (Anthropic).

"We ask the US government to support international efforts to develop the technical and governance tools needed to deliberately restrain the pace of automated AI development," the "Pacing the

Frontier" appeal states.

The authors worry that automated AI research is accelerating progress faster than the industry can build safeguards. The examples are telling: according to Anthropic, the company generates 80% of its code with Claude Code; OpenAI cut end-to-end serving costs by 20% using its Sol model; Kimi K3 designed its own chip for a nano model based on its own architecture. In other words, AI is already actively building the next generation of AI.

What does it mean?

Three letters in five days expose a deep split in the industry: some bet on openness and competition, others demand an international slowdown, and Anthropic opts for selective regulation. The 2026 regulatory season for AI has entered its hot phase, and the stakes are high — the question is who will control the most influential technology of the decade, and how.

Frequently asked questions

What are open-weight AI models?

Open-weight models are AI systems whose parameters are publicly available: they can be downloaded, studied, run locally and fine-tuned. Unlike closed models (GPT-5.6, Claude), their behavior is transparent to independent researchers, which, according to Microsoft and NVIDIA, is itself a safeguard against misuse.

How does

Anthropic's position differ from the Microsoft letter?

Microsoft and its 234 co-signatories consider open-weight models a less monopolistic and more transparent risk than closed ones. Anthropic is not opposed to open models in general, but demands limits on industrial-scale distillation and favors international regulation of the pace of AI development — a position that matches "Pacing the Frontier."