Bernie Sanders Proposes To Seize Half-Ownership of Large AI Firms

'A one-time 50 percent tax' inspired by a Columbia law journal article

Senator Bernie Sanders, socialist of Vermont, as seen on the cover of his 2023 book.

Senator Bernie Sanders, the socialist from Vermont who with Rep. Alexandria Ocasio-Cortez and Mayor Zohran Mamdani is an agenda-setter on the activist left wing of the Democratic Party, is proposing that the U.S. government seize 50 percent ownership of large artificial intelligence companies.

“I will soon be introducing the American A.I. Sovereign Wealth Fund Act. This legislation would give the public a direct ownership stake in the largest A.I. companies in our country,” Sanders announced in a New York Times opinion article published online. “It would create a sovereign wealth fund through a one-time 50 percent tax — not on the profits of OpenAI, Anthropic, xAI and other companies, but paid with something far more valuable than that: the stock.”

The proposal by Sanders—along with a May 27, 2026, piece in Time by Senator Elizabeth Warren, a Democrat from Massachusetts, “Why We Need to Tax AI”—signals that Washington politicians are scrambling for a share of the AI riches just as ardently as any investment banker hoping to lead an initial public offering.

With Democrats in the minority in both houses of Congress, for now the initiatives, like much activity by Sanders and his progressive allies, are more political posturing than any practical threat. An “Artificial Intelligence Data Center Moratorium Act“ Sanders introduced in March has attracted not a single Senate cosponsor. While Ocasio-Cortez was included on a Sanders press release announcing the legislation to “enact a reasonable pause to the development of AI to ensure the safety of humanity,” she has not yet introduced any version of it in the House, according to Congress’s legislative database.

Yet the legislation and opinion articles are good indications of the impulses and values that would guide Democrats if they do take over Congress following the midterm elections.

The Sanders plan is also a good case study of how universities and foundations are feeding left-wing ideas into the political system. Sanders acknowledges that he got the idea from two law professors, Jeremy Bearer-Friend of George Washington University Law School and Sarah Polcz of UC Davis School of Law. Bearer-Friend is a former aide to Warren, who herself is a former law professor. He’s also written for the Roosevelt Institute, a think tank funded by the Hewlett Foundation, George Soros’s Open Society Foundations, and the Ford Foundation that published a Bearer-Friend paper proposing a $1 trillion tax to fund racial reparations. (”As with any new trillion-dollar tax, we should expect litigation. But the power to tax has been a part of our Constitution since our founding and is not on its face legally dubious.”)

Bearer-Friend and Polcz published their article on December 30, 2025, in the Columbia Journal of Tax Law, proposing, “a unique in-kind tax payment structure that would require firms with ownership of AI to remit equity shares to the public.” Their article doesn’t specify a rate. “The actual rate to be adopted will largely be a political question,” it concedes. It does say the ideal rate “will provide the adequate level of public voice in corporate decisions to address public harms of AI” and “will provide a sufficient profit interest to compensate injured workers and creators,” while only “minimally” crowding out private investment.

In April, Polcz posted to social media a version of the law journal article that was adopted for The Hill, a Capitol Hill publication that’s a frequent fallback resort for opinion pieces written by academics that have been rejected by higher-profile outlets. “Know any lawmakers looking to tax AI? We’ve got a proposal for them,” she said, with remarkable candor about the motivation of the legislators.

At the moment the tax-AI urge seems mostly motivated by sentiment attributed to bank-robber Willie Sutton when asked why he targeted banks: “Because that’s where the money is.” Bloomberg News reports 19 new AI billionaires in the past year, worth a combined $59.3 billion. Anthropic, parent of the Claude AI tool, on June 1, 2026, announced it had “confidentially submitted a draft registration statement on Form S-1 to the U.S. Securities and Exchange Commission for a proposed initial public offering of our common stock.” On May 28, 2026, Anthropic announced it had raised $65 billion at a $965 billion post-money valuation.

The Columbia Journal of Tax Law article describes two goals: “compensating individuals for stolen data and workforce displacement.”

The irony here is that the individuals or companies whose data have been “stolen” are pretty well off. New York Times publisher A.G. Sulzberger, a fifth-generation heir to the Ochs-Sulzberger newspaper fortune, gave a speech on June 1 about what he called “original sin”—”a brazen theft of intellectual property that has occurred at an unprecedented scale. Tech giants strip-mine news websites without permission or compensation. They repackage these stolen goods as their own, siphoning off the audiences and revenue that otherwise would go to the news organizations that created this work. And this happens not just once during the training process, but countless times every single day.”

Sulzberger explained, “the Times took OpenA.I., its partner, Microsoft, and later Perplexity, to court, for brazen violations of our intellectual property rights … But lawsuits are slow and expensive — ours has already stretched two-and-a-half years and cost over 20 million dollars.” It’s far less “slow and expensive” to just get a friendly senator to seize the AI company’s stock and use those funds to pay off the newspaper companies or book publishers (which, conveniently enough, also are one of the few industries able to pay off sitting politicians directly via book “advances” without triggering an “ethics” stir).

As for the “workforce displacement,” newly laid-off tech-company workers may not be quite as prosperous as A.G. Sulzberger, but they were doing okay, too. The economy that gives them the best chance of finding new jobs is one that involves strong property rights and private-sector-driven growth, not arbitrary half-seizures of successful enterprises. I sent Professor Polcz a note asking what about AI is so exceptional to justify the 50 percent equity tax for workforce displacement. Why not a 50 percent tax on electric vehicles to compensate displaced gas-station attendants? Or a 50 percent tax on word-processing software to compensate displaced typewriter repairmen? A 50 percent tax on solar panels and wind turbines to compensate displaced coal and oil workers? What is the distinguishing feature here, if any? I didn’t hear back from her by deadline. The risk is similar to what the U.N. did when it created an agency to support Arab “refugees” after Israel’s creation—the existence of compensation winds up creating incentives to identify as displaced over extended timespans, rather than resettling.

Sanders is forthright about what he sees as the potential to use the AI money to pursue big-government welfare spending goals he’s been lusting after for years: “The proceeds would be used to ensure that every man, woman and child in our country has a decent and dignified standard of living, including health care, education and housing.”

Artificial intelligence has potential to drive efficiencies and quality improvements in health care, education, and housing. Imagine how robot construction workers could reduce housing costs, or self-driving trucks could speed and reduce costs for delivery of construction supplies. Yet Sanders seems to see only downside. “Left unchecked, it is likely that millions of truck drivers, bus drivers, taxi drivers and rideshare drivers will lose their jobs in the next decade,” he wrote in an April 16, 2026, Fox News opinion piece. “Why would any business want to hire a human worker when it can install AI and robotics and cut its labor costs by 80 to 90%? AI and robots don’t take a salary, need a vacation, require health care, or form a union. They just keep working — 24/7.”

The Columbia paper, in a footnote, claims “environmental devastation, national security risks, and tyrannical state surveillance are all noted harms of the growth of AI.” It also insists, “by positioning the state as a shareholder in private enterprise the proposal further expands capitalist systems rather than substituting them with Communist ones.”

It adds up to a dystopian view rather than a techno-optimist one. Yet if there’s an upside here, it’s that the forces of technology and free enterprise are so strong and fast-moving that even Sanders may be adjusting to a new reality. Less than three months ago he wanted to “pause” the development of AI Now he wants the government to own half of it. That half is worth a lot more without a data center moratorium than with one.

And at least, the Sanders proposal of “a one-time 50 percent tax” to confiscate a direct ownership stake may provide Republicans with a ready comeback the next time anyone suggests that the Trump’s administration’s “golden share”-style stakes in U.S. Steel or Intel are unprecedented authoritarian shakedowns.

READ MORE: Bernie Sanders Proposes To Seize Half-Ownership of Large AI Firms