The USA Wants AI Companies To Give Up Equity To A Public Wealth Fund

The United States is moving toward creating a sovereign wealth fund backed by direct equity stakes in its largest AI companies. Senator Bernie Sanders introduced the American AI Sovereign Wealth Fund Act in early June. The bill would require the largest AI firms to transfer 50% of their equity to a public fund through a one-time stock tax, held and managed for public benefit going forward.

A separate, smaller proposal is already on the table. Sam Altman has offered to donate roughly 5% of OpenAI to a similar public fund structure - worth about $42 billion at the company's current valuation of $852 billion. Altman has floated extending that model to Anthropic, Google, and Meta.

The government has actually done something like this before, though not with an AI company.

The Trump administration gave Intel, the computer chip maker, several billion dollars to help build new chip factories in the US. In exchange, the government didn't just hand over the money for free. It took a 10% ownership stake in Intel itself, similar to how a shareholder owns part of a company. So if Intel stock goes up, the government's stake is worth more too.

This is the closest example we have of what a government stake in an AI company might actually look like in practice.

The logic behind both AI fund proposals - Altman's 5% equity stake and Sanders' 50% - is the same. AI is being built on decades of public research, data, and infrastructure, and proponents argue the wealth it generates shouldn't flow only to private shareholders and venture capital.

The reference point in both cases is the Alaska Permanent Fund or Norway's petroleum fund: state ownership of a strategic asset, invested in markets, with returns paid back to citizens.

As of today, neither proposal has passed. Sanders' bill remains in committee. Altman's offer is a public statement, not a signed agreement.

But a Verasight poll shows 69% of American workers are in favour of some version of public equity ownership in AI firms. And that support is what's keeping both proposals alive in Washington despite no legislative movement yet.

If the 50% version becomes law, the government would instantly become one of the biggest owners in the entire AI industry. It would be sitting at the same table as venture capital firms, big tech company boards, and the sovereign wealth funds that Gulf and Asian countries already run. Abu Dhabi already has a fund called MGX with $49 billion just for AI investments.

A US government fund at the scale Bernie Sanders has proposed wouldn't just quietly own stock and collect dividends. It would get a real vote in company decisions. And it would have influence over how future AI projects - data centers, chip factories, power plants - actually get financed.

A one-time 50% tax on company equity is a much bigger move than Altman's voluntary 5% offer, and it changes the deal for anyone building or investing in an AI company going forward.

Part of the reason the US has stayed ahead in AI is that the government has mostly stayed out of owning pieces of these companies. If Congress can take half of an AI company's equity by passing a law, that's a new kind of risk investors have to think about. And it wouldn't necessarily stop at AI. The same idea could get applied to biotech, energy, or defence companies down the road.

Even if these specific bills don't pass, the pressure behind them likely isn't going away. People are frustrated that a small number of companies and investors are capturing most of the wealth from AI while jobs get displaced by the same technology.

Other governments have already gotten comfortable owning stakes in AI infrastructure. This isn't a new idea in other parts of the world. What's changed is that Washington is now having this conversation in public instead of avoiding it.

And that shift - from a fringe idea to an actual proposal in Congress - is the part worth watching closely, no matter which version of the fund eventually gets built.