Not a ban. A share.

We argue for one thing: that the public should hold a substantial ownership stake in the companies building frontier AI — OpenAI, Anthropic, Meta, xAI and their peers — through the US federal government, and through state-level stakes alongside it.

This is an economic argument, not a cultural one. These companies are building systems that are already reshaping whole categories of work — including ours. The value they create is built in large part on work the public already produced: the writing, images, code and recordings the models learned from. When infrastructure this consequential is owned narrowly, the returns are owned narrowly too. A public stake is a claim on that upside on behalf of the people whose work is in the training data and whose jobs are being redrawn.

Public ownership stakes in strategically important companies are not novel. Governments hold equity through sovereign wealth funds and take positions in firms they consider critical. We are asking for that ordinary tool to be pointed at an extraordinary industry.

The strongest argument against us

The best objection is that government ownership could slow the work, politicize it, or capture it — that a stake becomes a lever, and the lever gets pulled for reasons that have nothing to do with good technology. It is a real risk and we take it seriously. Our answer is that a stake is not the same as control: ownership can be structured for a share of the returns and a seat at the table without a veto over the roadmap, the way a passive investor holds equity without running the company. The design of that structure is exactly the debate worth having.

What we are not arguing: we are not against these tools — the people here use them every day. We are not naming a party, a candidate, or an election; this is about ownership, on the merits, whoever is in office. And we are not calling the companies villains. The question is who owns the upside, not whether the work should happen.