Sanders' AI Wealth Fund Proposal Amounts to State Bailout of Tech Giants, Critics Argue

The state would sit as a junior partner while workers owned nothing and decided nothing.
Critics argue Sanders' proposal leaves AI companies in private hands despite government equity stakes.

As artificial intelligence reshapes the labor landscape and concentrates extraordinary power in a handful of private hands, Senator Bernie Sanders has proposed a sovereign wealth fund that would grant the federal government a 50 percent equity stake in the dominant AI corporations. The swift convergence of support from the Trump administration, AI executives, and Sanders himself has prompted critics to ask whether the proposal represents a genuine redistribution of power or a sophisticated mechanism for legitimizing monopoly, absorbing public discontent, and binding transformative technology to military ends. At stake is not merely a tax structure but a fundamental question about whether the collective intellectual inheritance of humanity — the books, art, and labor from which these systems were built — can be reclaimed through a partnership with the very state and capital that extracted it.

  • Nearly 400,000 American jobs have been cut in five months, with AI named as the primary cause, and the pressure for a political response has become impossible to ignore.
  • Within days of Sanders' op-ed, Trump signed an executive order, Altman visited Sanders' office, and all three parties converged on the same public-private framework — a alignment critics call a warning sign, not a victory.
  • The proposal would leave AI companies privately owned, profit-driven, and executive-controlled, with the government holding shares but workers holding nothing — a junior partnership dressed in the language of reclamation.
  • AI firms are racing toward IPOs at valuations exceeding $850 billion, and a government holding half their stock would have every incentive to prop those valuations up rather than challenge the industry's power.
  • Anthropic's unrestricted AI model is already deployed in collaboration with the US government and was used to generate over 1,000 bombing targets on the opening day of the US-Israeli assault on Iran, illustrating what 'state direction of AI' looks like in practice.
  • Critics argue Sanders is performing his established role — absorbing working-class anger and redirecting it into the Democratic Party — rather than advancing a path toward genuine democratic control of technology.

Senator Bernie Sanders published an op-ed proposing that the federal government acquire a 50 percent stake in the major AI companies — OpenAI, Anthropic, xAI, and others — through a one-time tax paid in company stock, creating what he calls an American AI Sovereign Wealth Fund. He framed it as a way to reclaim technology built from humanity's collective intellectual inheritance: books, art, journalism, and research seized by billionaires without permission or compensation.

The response was revealing. Within a day, Trump signed an executive order granting the government early access to the most powerful AI models for national security purposes — and Sanders endorsed it. Days later, Trump disclosed his administration had spent over a year negotiating its own equity stake in OpenAI, calling it 'almost a partnership with the American public.' Sam Altman visited Sanders' office and expressed support for public ownership, objecting only to the 50 percent figure. The convergence of the Trump administration, the AI executives, and Sanders on a single framework struck critics not as a coalition of the unlikely but as a coalition of the aligned.

Under the proposal, the companies would remain privately owned and operated by the same executives, driven by the same profit motive. The state would hold voting shares and board seats but workers would own nothing and control nothing. No legislation has been introduced; Sanders acknowledged the mechanics would emerge 'in the coming weeks.' Critics argue the scheme functions primarily as a framework for future corporate bailouts — the AI companies have staked hundreds of billions on valuations that may never materialize, and a government holding half their shares would have a vested interest in guaranteeing those valuations against collapse.

Sanders invoked Norway's sovereign wealth fund as a model of shared prosperity. Critics counter that the fund is invested entirely abroad, has functioned as an instrument of austerity, holds stakes in Israeli military contractors, and is overseen by a former NATO secretary-general whose government crushed an oil workers' strike on its first day.

Trump's executive order, which Sanders welcomed, commits the government to deploying AI rapidly for national security. Anthropic has already released an unrestricted model in collaboration with the US government; embedded in Palantir's Maven system, it was used to generate over 1,000 bombing targets on the opening day of the US-Israeli assault on Iran. The same systems operate in Gaza, power mass surveillance, and enable censorship of anti-war dissent.

For critics, Sanders' proposal follows the pattern of his career: channeling working-class anger back into the Democratic Party and the capitalist framework rather than toward independent political organization. The question of who controls AI, they argue, cannot be settled by negotiating tax terms with Sam Altman. Technology built from the accumulated labor and knowledge of humanity must become the common property of the international working class — expropriated from the corporations and placed under democratic worker control.

Senator Bernie Sanders published an op-ed in the New York Times last week proposing that the federal government acquire a 50 percent stake in the major artificial intelligence companies—OpenAI, Anthropic, xAI, and others—through what he calls an American AI Sovereign Wealth Fund Act. He framed the measure as a way to wrest control of the technology from the small group of billionaires who currently dominate it. The response that followed revealed, according to critics, the true character of the proposal.

Within a day of the op-ed's publication, President Trump signed an executive order granting the government early access to these companies' most powerful AI models for national security review. Sanders endorsed the order. Two days later, aboard Air Force One, Trump disclosed that his administration had spent more than a year negotiating its own equity stake in OpenAI, describing it as "almost a partnership with the American public." When asked about Sanders' proposal, Trump suggested that the economic views of his supporters and Sanders' supporters "aren't that far apart." Meanwhile, OpenAI CEO Sam Altman met with Sanders in the senator's office and expressed support for public equity ownership in AI companies, objecting only to the 50 percent figure. The alignment appeared unmistakable: the Trump administration, the AI executives, and Sanders were converging on a single framework.

Sanders' case rests on real grievances. In the first five months of 2026, American employers announced nearly 400,000 job cuts, with AI cited as the primary driver. Sanders argues that AI has been constructed from humanity's collective intellectual inheritance—books, songs, art, journalism, code, research—which the tech oligarchs seized without permission, acknowledgment, or compensation. "It's time for us to reclaim it," he wrote. The diagnosis is sound. The remedy, critics contend, is not.

Under Sanders' proposal, the government would impose a one-time 50 percent tax on the AI giants, collected not in cash but in company stock. Through voting shares and equal board representation, the government would gain the power to "block decisions that hurt our citizens." Yet the companies would remain privately owned and operated by the same executives and investors, driven by the same profit motive. The state would sit as a junior partner in a public-private arrangement while workers owned nothing, controlled nothing, and decided nothing. No bill has been introduced; Sanders acknowledged that "the specific spending priorities and the mechanics of implementation" would emerge only "in the coming weeks."

Critics argue the scheme functions primarily as a framework for future corporate bailouts. The AI companies have invested hundreds of billions of dollars in data centers based on valuations that may never materialize. Anthropic filed confidentially for an initial public offering on the day Sanders' op-ed appeared; OpenAI, valued above $850 billion, is preparing its own for as early as September. A government holding half the shares would acquire a vested interest in propping up these valuations, guaranteeing against collapse while providing democratic legitimacy as mass layoffs turn public opinion against the industry.

Sanders invoked Norway's $2 trillion sovereign wealth fund as a model, claiming it embodies the principle that oil wealth "should be used to improve life for all of its people." The reality, critics say, is inverted. The fund is invested entirely abroad; Norwegians own and control none of it. It has functioned as an instrument of austerity, capping annual budget allocations at 3 to 4 percent of its value. Overseas, it holds close to 1.5 percent of all listed shares globally and is invested in dozens of Israeli companies involved in military operations. It is overseen by Jens Stoltenberg, a former NATO secretary-general, and in 2022 the Norwegian government that runs it crushed an oil and gas workers' strike on its first day to maintain energy supplies to a Europe at war with Russia.

Trump's executive order, which Sanders welcomed, illuminates what state direction of AI means in practice. The order commits the government to "work closely with industry to ensure that the best and most secure technology is deployed rapidly to confront any and all threats to our country." Anthropic released its most powerful AI model this week in two versions: the public Claude Fable 5 and an unrestricted Claude Mythos 5, deployed through a program run "in collaboration with the US government." Washington has no intention of curbing the AI monopolies; it is harnessing them for military purposes. On the opening day of the US-Israeli assault on Iran, Anthropic's Claude, embedded in Palantir's Maven system, was used to generate more than 1,000 bombing targets. The same systems direct operations in Gaza, power mass surveillance, and enable censorship against anti-war and socialist opposition.

Sanders has not stumbled into the wrong answer by accident. He is an experienced political operative, and this scheme serves the function he has served throughout his career: channeling worker and youth anger back into the Democratic Party and the capitalist framework. In 2016, he campaigned on a "political revolution," won 13 million primary votes, then endorsed Hillary Clinton. He repeated the pattern in 2020 and was among the first to back Joe Biden in 2024. He operates as a pressure valve, releasing the building force of working-class discontent before it can take independent political form. The question of who controls AI cannot be settled by negotiating tax terms with Sam Altman or entrusted to the capitalist state. AI is the product of accumulated international labor and knowledge and must become the common property of the international working class, expropriated from the major technology corporations and transformed into public utilities under democratic worker control.

It's time for us to reclaim it
— Bernie Sanders, on reclaiming AI built from collective human knowledge
Almost becomes a partnership with the American public
— President Trump, describing his administration's proposed stake in OpenAI
The Hearth Conversation Another angle on the story
Inventor

Why does Sanders' proposal look like a bailout rather than a genuine check on corporate power?

Model

Because the companies stay private, stay profitable, and stay in the hands of the same people running them now. The government becomes a shareholder, not an owner. It's a partnership, not a takeover.

Inventor

But wouldn't government board seats give workers some say in how AI gets used?

Model

The government sitting on a board isn't the same as workers controlling anything. The US government is already using AI to generate bombing targets and power surveillance. A state stake doesn't change the purpose; it legitimizes it.

Inventor

What about the job losses Sanders mentions—nearly 400,000 in five months?

Model

That's real and urgent. But Sanders' fund doesn't prevent a single layoff. It just gives the state a financial interest in keeping the companies afloat, which means keeping them profitable, which means continuing to replace workers with machines.

Inventor

He points to Norway's sovereign wealth fund as a model. Isn't that working?

Model

Norway's fund is invested entirely abroad. Norwegians don't own or control it. It enforces austerity at home, profits from Israeli military operations, and crushed a workers' strike to keep gas flowing to a war. That's the actual precedent.

Inventor

So what would genuine control look like?

Model

Expropriation. The technology belongs to the workers who built it. They should own it collectively, control it democratically, and decide what it's used for. Not through a government fund, but through the workers themselves.

Inventor

And Sanders knows this isn't what he's proposing?

Model

He's a seasoned politician. This serves his real function: keeping worker anger inside the Democratic Party instead of letting it become something independent and genuinely threatening to the system.

Contact Us FAQ