Sanders proposes public ownership of 50% of OpenAI, Anthropic, and xAI

Senator Bernie Sanders (I-Vermont) has introduced the most ambitious federal legislation to date on public ownership of artificial intelligence: the American AI Sovereign Wealth Fund Act. Co-sponsored by Representative Ro Khanna (D-California), the bill proposes a one-time 50% tax on shares of the largest AI companies, which would be deposited into a sovereign wealth fund managed by the government for the benefit of all American citizens.

What exactly does it propose?

The central mechanism is a capital tax, not a profit tax. AI companies with annual revenues exceeding $200 million would have to cede 50% of their shares to the fund. Affected companies would include OpenAI (valued at $852 billion for its confidential IPO), Anthropic (valued at approximately $1 trillion), xAI, and likely the AI divisions of Microsoft and Google, depending on how the threshold is defined. The resulting sovereign wealth fund would be valued at around $7 trillion.

An Independent Commission for Democratic AI, composed of seven members appointed by the president and confirmed by the Senate, would manage the shares. It would have two explicit mandates: block business decisions that harm citizens and promote policies that benefit them. The commission would have equivalent representation on the boards of each participating company.

Why is this important?

This proposal represents a radical shift in AI governance. Sanders argues that AI is built on the collective knowledge of humanity and the creative work of tens of millions of people, so its benefits should be shared. The model is inspired by Norway's Government Pension Fund, funded by oil revenues and valued at over $2 trillion.

If approved, it would be the first time the U.S. federal government holds a direct stake in large-scale private tech companies. The generated income would be distributed as annual direct payments: Sanders calculates over $1,000 per American adult, with growth prospects as AI generates more economic value.

The historical context is key. Since the 1980s, the U.S. has favored a free-market approach to technological innovation, with tax breaks and little direct intervention. However, the 2008 financial crisis and the bailout of banks like Citigroup (where the government temporarily acquired stakes) set a precedent for state intervention in extreme cases. More recently, the CHIPS and Science Act of 2022 injected $52 billion into semiconductor manufacturing, but without taking ownership. This bill goes much further: it's not about subsidies or loans, but a partial expropriation of major AI players.

Compared to other global models, Norway is the closest example, but its fund is financed by natural resources, not tech stocks. In China, the government holds stakes in companies like Baidu and Alibaba through state entities, but not via a one-time tax. The European Union, for its part, proposed a "robot tax" in 2017 and more recently a digital solidarity fund, but neither has succeeded. This U.S. bill would be the most aggressive in terms of direct public ownership.

Consequences and challenges

The proposal faces enormous political hurdles. Even within the Democratic Party, there is division over state intervention in the tech industry. Republicans will likely reject it as expropriation. Additionally, affected companies could go to court arguing that the capital tax violates the Fifth Amendment (taking of property without just compensation).

If implemented, the sovereign wealth fund would give the government unprecedented power to influence strategic decisions of major AI companies, from model development to safety policies. It would also create an incentive for the government to maximize the fund's value, potentially leading to conflicts of interest between regulation and profitability.

An economic impact analysis suggests the fund could generate between $300 billion and $500 billion annually in dividends, assuming a 5-7% return on the $7 trillion. However, company valuations could plummet if the market interprets the measure as a sign of regulatory instability. For example, after the announcement, shares of tech companies not directly affected (like Meta or Amazon) fell slightly in pre-market trading, reflecting fears of a contagion effect.

For AI startups, the message is clear: scaling too fast could expose them to future expropriation. This could slow investment in the sector or redirect it to more favorable jurisdictions, such as Singapore or the United Arab Emirates, which have already created their own tech sovereign wealth funds. In the long term, the proposal could fragment the global AI ecosystem, with U.S. companies seeking corporate structures abroad to avoid the tax.

What should readers know?

For now, it's just a legislative proposal with little chance of passing in the short term. However, it reflects a growing global debate on how to distribute the benefits of AI. Countries like the European Union and China are already exploring models of public ownership or tech sovereign wealth funds. This bill could serve as a catalyst for broader discussions on democratizing AI.

For investors and startups, the mere threat of such a measure could accelerate the relocation of AI companies to more favorable jurisdictions. For citizens, it opens the possibility of a universal dividend funded by technology.

In summary, the American AI Sovereign Wealth Fund Act is a milestone in AI policy, but its viability is uncertain. What is clear is that the conversation about who benefits from AI has taken a radical turn, and this bill, though unlikely to become law, is already influencing public debate and the strategies of tech companies.