OpenAI has reportedly floated an offer to give the federal government a 5 percent ownership stake in the company. The Financial Times reported this week that CEO Sam Altman raised the idea with President Trump, Commerce Secretary Howard Lutnick, and Treasury Secretary Scott Bessent. Altman has also suggested that other leading American AI developers contribute comparable equity to a public investment vehicle modeled on the Alaska Permanent Fund. At OpenAI’s $852 billion valuation from its March funding round, a 5 percent stake in OpenAI alone would be worth about $42.6 billion. Altman argues this is the cleanest way to let the public share in AI's upside.

The idea is still preliminary and might need Congress to approve. But the predictable attacks have already started. Some on the right call it socialism, while some on the left think it doesn’t go far enough. Both reactions miss the mark. Done right, this is sound public finance , and perhaps the most workable version of an American sovereign wealth fund yet proposed.

AI is often presented as a triumph of private ingenuity. It is that, but not only that. Today’s large language models were trained on books, journalism, music, code, art, and scientific papers created by millions of people who never sat at any negotiating table and never received a market price for their contribution. Long before venture capital poured money into AI, decades of federal research also helped build the computing and machine-learning advances that made these systems possible.

When public inputs help create extraordinary private returns, the public has a legitimate claim to part of the upside. Senator Bernie Sanders has seized on that point with his American A.I. Sovereign Wealth Fund Act , which would require large AI companies to pay a stock tax equal to half their equity. Sanders identifies a real free-rider problem, even if his solution goes much too far. A 50 percent equity claim would scare private capital away from firms that are still unprofitable and still need outside financing.

The OpenAI offer avoids that mistake. A fixed, known 5 percent dilution is something investors can simply build into the price they pay for shares. Remove that uncertainty and capital keeps flowing.

Sovereign Wealth Funds Solve Real Problems

Sovereign wealth funds are useful because they respond to two failures at once. Private markets chronically undersave, leaving future generations with less capital than a farsighted society would choose to bequeath. Governments, meanwhile, run on election cycles. Elected officials have every incentive to spend windfalls while they are still in office, then leave the consequences for someone else. A well-built sovereign wealth fund converts a temporary source of wealth into a portfolio of long-run productive assets, and it gives the state a profit-and-loss measure of success that ordinary spending programs lack.

Norway's Government Pension Fund Global holds more than $2 trillion in assets and returned 15.1 percent in 2025. Abu Dhabi's sovereign funds manage roughly $1.7 trillion, including Mubadala, which has returned more than 10 percent annually over the past five years. Those returns are in the range private asset managers would be proud of. Nor is any of this an exotic foreign practice. More than 20 states, including Alaska, Texas, and North Dakota, already run public wealth funds.

The United States also has another problem that a fund could help solve. Because the dollar serves as the world’s reserve currency, Washington borrows far more cheaply than its fiscal behavior deserves, and it spends the proceeds almost entirely on consumption. A national fund could redirect at least a portion of that privileged financing into assets that earn income for the Treasury instead of simply adding to the Treasury's interest bill.

The Alaska Model Is the Right Template

Altman’s choice of the Alaska Permanent Fund is no accident. Alaska created the fund in 1976 to turn a finite oil windfall into a lasting income stream. The fund now holds roughly $91 billion and has paid every eligible Alaskan a dividend every year since 1982, including a $1,000 payment last year. The oil will eventually run out. The fund can keep earning compounding returns.

The AI boom has the character of an oil strike. Nobody knows how long today’s extraordinary valuations will last or how the industry will eventually consolidate. That uncertainty is a reason to take some of the upside now, not to wait. Moreover, because the equity would be contributed rather than purchased, taxpayers also face none of the usual risk of buying into a bubble at the top. If valuations fall, the public has lost nothing. If valuations rise, the public shares in the gain.

With a payout rule similar to Alaska’s, which draws about 5 percent of fund value annually, OpenAI’s stake alone could support roughly $2 billion a year in distributions. If the same arrangement reached the other leading labs, the annual flow would be several times larger. Congress could use the money to reduce deficits, fund tax relief, or pay dividends directly to American households.

Alaska also demonstrates that the fear that dividends automatically create dependency or invite politicians to raid the principal is overblown. The program is successful precisely because every resident can see the connection between sound fund management and the check that shows up once a year in the mailbox. That ownership stake gives ordinary voters a reason to defend the fund. It creates a healthy political feedback loop, and it has helped the fund survive five decades of budget fights and oil busts.

Some free-market advocates object that any federal stake in a private company is nationalization by another name. But the objection misses why government activities usually fail. The trouble is that they typically operate outside the discipline of the price system, with no profit-and-loss signal to reveal when value is destroyed. An equity stake avoids that trap entirely. The government would hold shares whose worth is set by the market. Still, any stake should be structured passively, with no voting rights and no board seats, following the precedent of the Intel arrangement , in which the government took a roughly 10 percent nonvoting position last year.

Washington is already in the equity business. Since January 2025, the federal government has taken $26.7 billion in equity and similar positions across thirty deals, including in semiconductors, rare earths, and quantum computing. Public ownership already exists. The question is whether Congress imposes clear rules on these deals before ad hoc dealmaking hardens into a permanent industrial policy run by the executive branch alone.

The conservative case for a fund is that it should be transparent, independent, and a stabilizing force on the federal budget. Properly designed, a fund can address chronic undersaving , and it gives Congress a visible opportunity cost. Once the fund exists, every program should implicitly answer the question of whether it delivers more value than simply leaving the money invested. A fund also puts idle public assets to work earning returns. That is not socialism. It is government participation in market capitalism.

President Trump has taken equity stakes in strategic companies and signed an executive order directing his administration to draw up plans for a sovereign wealth fund. Senator Sanders wants public ownership of AI on a much larger scale. Now the best-known AI executive in the world is volunteering his own company’s shares. When the populist right, the socialist left, and Silicon Valley are all circling around the same idea, it is time to take the idea seriously.

AI companies have practical reasons to make a deal. Washington’s scrutiny has sharpened in recent weeks, with federal officials intervening in model releases on national-security grounds and state attorneys general opening investigations . The industry needs legitimacy. A modest equity contribution could buy durable public support at a time when public patience with Silicon Valley is thinning. The leading labs are also preparing for stock market listings, and settling the question of the government's role before their shares trade would remove a major overhang for prospective investors.

American workers are understandably nervous about a technology built to automate tasks now done by people. Some AI leaders also warn in apocalyptic terms about the technology’s risks. Against that backdrop, telling workers that the gains will trickle down to them someday is not enough. Ownership is a more credible answer. If AI creates the wealth its builders promise, Americans should hold a direct claim on part of it.

Congress and the president should accept Altman’s offer, write careful rules around sovereign wealth fund governance, and create an institution that can outlast the current AI hype cycle. The AI industry is offering Americans equity in their own future. It would be a dereliction of duty to refuse.