Bernie Sanders proposes 50% stock tax on AI labs to fund American AI Sovereign Wealth Fund

Bernie Sanders Proposes 50% Stock Tax on AI Labs

Senator Bernie Sanders is introducing the American AI Sovereign Wealth Fund Act, which would impose a one-time 50% tax on AI company stock, with proceeds funding a government-controlled sovereign wealth fund. The fund would grant the federal government voting rights, board seats, and the ability to pay dividends directly to American citizens.

The proposal rests on a straightforward moral argument: AI is built on humanity's collective knowledge, so the wealth it generates should benefit the public, not individual billionaires like Elon Musk or Sam Altman. Sanders published a guest essay in The New York Times laying out the case.

The mechanics create obvious complications. A 50% equity stake would give government control, but the dividend timeline is murky. Most tech companies take decades to return cash to shareholders, preferring reinvestment. The hosts note this tension without resolution in the transcript.

The scope problem is thornier. Sanders has not clarified which companies qualify as "AI labs." Google, Microsoft, Meta, and Amazon all have significant AI operations but are not pure-play AI companies. Intel and Nvidia make chips, not models. Amazon's Rufus and Walmart's Sparky are AI-powered retail assistants but rely on foundation models built elsewhere. If the bar is simply "anything built on humanity's collective knowledge," the logic extends to nearly all enterprise software, infrastructure, and consumer products. One host jokes that by that standard, the government could claim half of everything—nails, roads, combustion engines.

The internal contradiction is sharp. Dean Ball, founder of Hyperdimensional, notes Sanders is simultaneously pursuing two incompatible positions: banning data centers through a moratorium act while proposing to take ownership stakes in AI labs. Ball frames this as evidence that Sanders' AI safety concerns are performative cover for a capital redistribution agenda.

One counterargument surfaces briefly: if Sanders genuinely believes AI is an existential risk, government board control over every major lab could enable coordinated slowdowns. But the hosts suggest a regulatory approach—an AI FDA—would accomplish the same goal without equity ownership.

A contrarian read suggests the proposal may be accidentally bullish. Sanders, not typically associated with capitalist wealth creation, is implicitly valuing AI companies as so economically disruptive that a 50% public stake makes sense. The data center moratorium may be performative; the real bet is that AI stocks are undervalued relative to their economic impact.

This remains an op-ed, not legislation. It signals Sanders' positioning within an AI policy conversation dominated by different camps—safety advocates, accelerationists, and redistributionists. How the proposal evolves, if at all, will depend on which concerns gain legislative traction.