Tech CEOs push Uncle Sam on open-weight AI, but Meta and Google skip the memo

A coalition of tech leaders just fired off a letter to the US government making the case for open-weight AI models — but the signatory list is almost as interesting as who’s missing from it. The group, which includes executives from stability-focused AI firms and infrastructure players, is trying to educate policymakers on what open-weight actually means and why it matters for national competitiveness. The letter lands at a moment when Washington is increasingly skittish about powerful AI models circulating without guardrails.

Here’s the thing about open-weight: it’s not open-source in the traditional sense. You get the trained parameters, not the training data or the full recipe. That distinction is exactly what these execs are trying to drill into lawmakers’ heads before they craft regulations that treat everything like a nuclear launch code. The argument goes that open-weight models let researchers and startups build on existing work without reinventing the wheel — and without handing foreign adversaries a blueprint.

The elephant not in the room? Meta and Google. Neither company put ink on this letter, which raises an obvious question. Meta has been the poster child for open-weight releases with its Llama models. Zuck’s whole pitch has been that open ecosystems beat closed ones. So why sit this one out? Could be political calculus, could be internal disagreement about how hard to push, or could be that Meta doesn’t want government attention trained on its release practices.

What this really signals is a splintering in how the industry wants to handle regulation. Some players want to be at the table shaping the rules; others seem to think the safest move is staying quiet and hoping lawmakers don’t look too closely at what they’re shipping. The letter represents a bet that engagement beats avoidance. Whether Congress actually listens — or just files it alongside every other tech white paper gathering dust in a Capitol Hill office — remains the open question.