Twenty-five technology companies published a letter on July 24 titled “Open Weights and American AI Leadership,” urging United States policymakers not to place premature restrictions on so-called open-weight AI models, which anyone can download and run.
The signatories included Nvidia, Microsoft, Meta, IBM, Hugging Face, Palantir, Mozilla and the Linux Foundation. Notably missing were the biggest closed-model developers, OpenAI, Anthropic, Alphabet and Amazon.
OpenAI added its signature only later, after its absence was widely noticed, and chief executive Sam Altman reposted a note on X backing open-source models. Anthropic declined to sign, with its chief executive Dario Amodei having previously dismissed open source as “a red herring.”
The debate has become political. David Sacks, the White House AI adviser, accused the closed-model companies of “lobbying to ban open-source rivals,” while officials have signalled they see no need to restrict open-source AI for now.
Part of the pressure comes from China. Moonshot AI’s Kimi K3, a 2.8-trillion-parameter open model released in mid-July, has benchmarked competitively with leading American systems while charging far less, at 3 dollars per million input tokens and 15 dollars per million output tokens. Separately, OpenAI and Anthropic have been aligning in Washington to warn officials about the risks of powerful Chinese open models.
Why it matters: open models are free to download and cheaper to run, which makes them especially valuable for the startups, researchers and developers in Africa who cannot always afford costly commercial AI subscriptions.
Comments 0