Can Sanders' AI equity mandate survive a Takings Clause challenge?

If S. 4825 ever moved from proposal to enforceable law, the threshold question for an affected AI company would not be whether the policy is unusually aggressive. It would be narrower and more useful: does Congress defeat a Fifth Amendment Takings Clause claim by calling a mandatory transfer of half the company’s equity an excise tax? This is a U.S. constitutional risk analysis for litigators and in-house counsel, not legal advice. The bill was introduced on June 18, 2026, and the available materials do not identify any Supreme Court case squarely deciding this tax/equity-transfer hybrid. [1]

The bill does not stop at collecting money

The American AI Sovereign Wealth Fund Act is described by its sponsor as a way for the public to own half of the largest AI companies. Sanders’ Senate release frames the bill as creating an AI sovereign wealth fund, while his op-ed states the more direct premise: “The Public Should Own Half of the Big A.I. Companies.” [1][2]

For Takings Clause purposes, the operative point is not the projected size of the fund or the distributive theory behind it. The important feature is the instrument Congress would use. The proposal is styled as an excise tax on “systemically important AI activity,” but the obligation would be satisfied through a transfer of equity rather than an ordinary cash payment. Available analysis of the bill describes a 50% equity transfer to a federal sovereign wealth fund, with coverage that reaches share classes and corporate governance arrangements rather than merely imposing a dollar-denominated tax liability. [3]

That is why the “tax” label cannot be the end of a serious analysis. A cash excise tax leaves the government as a collector of funds. This proposal would make the government a shareholder with voting power and governance leverage. The difference is not semantic; it changes who can vote, who can block transactions, and who occupies the boardroom.

Why stock and control rights make the Takings argument credible

A plaintiff’s cleanest Takings Clause theory would begin with the equity itself. Stock is not merely an accounting entry. It bundles claims to economic value with governance rights defined by corporate law and the company’s governing documents. If the statute requires shareholders to deliver voting stock to the federal government, the injury is not just that the company has less money. The injury is that ownership and control rights have been reassigned by federal command.

The Volokh Conspiracy analysis published at Reason supplies the central constitutional frame. It argues that the mandate resembles a per se taking because the government would appropriate a defined property interest, and it invokes Loretto v. Teleprompter Manhattan CATV Corp., the 1982 Supreme Court decision treating a permanent physical occupation as a taking without needing a balancing inquiry. [5]

That analogy does useful work, but it should not be overstated. Loretto was not a case about AI firms, federal stock mandates, sovereign wealth funds, or excise taxes. The argument is that the principle extends: when the government compels a permanent transfer of a discrete property interest, a court should look through the statutory packaging and treat the transfer as an appropriation. That is a serious argument, not a holding already on the books.

The governance features strengthen the analogy. A statute that took a slice of revenue might be analyzed as a tax. A statute that forced the transfer of voting shares, supplied board seats, preserved a federal voting block, and preempted corporate defenses would look less like revenue collection and more like compelled co-ownership. The Truth on the Market analysis emphasizes the control consequences, including the creation of a federal voting block with implications for mergers, takeovers, and other corporate-control transactions. [4]

The tax label helps the government, but it does not answer the property question

The government’s first move would be obvious: Congress has taxing power, and the statute calls the obligation an excise tax. Courts are generally cautious about constitutional theories that would turn every hard tax into a taking. If the only burden were a cash payment measured by AI activity, the Takings Clause claim would have a much steeper climb.

But the label does not erase what the statute makes the taxpayer hand over. The Reason analysis makes the substance-over-form point directly: a law that requires a company or its shareholders to convey equity to the government is not insulated from Takings Clause scrutiny merely because Congress routes the command through the Internal Revenue Code vocabulary. [5]

The strongest plaintiff version would be instrument-specific. It would not say that all in-kind taxes are takings or that Congress may never condition activity on a transfer of value. It would say that this particular statute identifies a valuable property interest — stock with voting and governance attributes — and transfers that interest to the sovereign without just compensation. That framing avoids a broad attack on the tax system and keeps the court focused on the unusual statutory mechanism.

The bylaw-preemption feature matters here. Corporate charters and bylaws often allocate control protections, voting thresholds, and defensive mechanisms. If federal law overrides those private and state-law arrangements to ensure the sovereign fund receives effective voting power, the statute is doing more than measuring tax liability. It is rearranging corporate governance so the government’s new equity position can function. The American Action Forum analysis flags that preemption issue as part of the bill’s mechanism. [3]

The government’s better defenses are not frivolous

A good complaint would not make the government’s defenses disappear. The taxing-power argument would be the most important one, but not the only one. The government would likely characterize the transfer as part of a comprehensive public-benefit regime aimed at allocating gains from systemically important AI activity. It could argue that the sovereign wealth fund’s dividends or public distributions supply a compensatory benefit, reducing or defeating the claim that private property has been taken without just compensation.

That defense would raise difficult fit problems. Just compensation ordinarily runs to the owner whose property is taken, not to the public at large. A dividend program for citizens or residents may be a political answer to the bill’s distributive purpose, but it is not automatically a constitutional answer to an AI company’s shareholders. Still, because no directly analogous Supreme Court case is identified in the current materials, counsel should expect the government to press the point aggressively rather than concede that a per se rule controls.

The government could also resist the property characterization. It might argue that firms engaged in systemically important AI activity operate in a heavily regulated field, that share ownership in such companies is already subject to extensive federal constraints, and that the statute defines a lawful condition for participation in a market with unusual public consequences. That answer should not be confused with a knockout. Regulation of an industry does not usually mean the government can take half the voting stock without compensation. But it would give the government a doctrinal path away from the plaintiff’s preferred per se-taking box.

There is also a remedial complication. A court might agree that the mandate burdens property interests while still hesitating over whether the claim should be treated as a per se taking, a regulatory taking, an unconstitutional condition, or a tax challenge with Takings Clause characteristics. The plaintiff wants Loretto-like simplicity. The statute’s hybrid design gives the government room to argue for a less categorical framework. [5]

A passive government stake would be a poor safe harbor

The Intel comparison is useful only if it is kept small. Current commentary treats the Trump administration’s 10% passive Intel stake, with no board seats, as a much less constitutionally aggressive reference point. [4]

That comparison does not validate a mandatory 50% AI equity transfer. It shows the opposite. A passive minority stake without board representation does not carry the same control consequences as a federal block of voting shares paired with board seats, veto leverage, bylaw preemption, and immunity protections. If a court were looking for the point at which government equity ownership becomes constitutionally different in kind, the Sanders proposal gives plaintiffs far more to work with than a passive industrial-policy investment.

What an affected company would need to map before suing

The proposal’s near-term enactment odds appear low in the current Republican-controlled Congress. That does not make the issue useless for counsel. The reason to map it now is not that S. 4825 is certain to pass. It is that the statutory concept — compulsory public ownership of AI firms through a tax-labeled equity mandate — may reappear in a narrower or more politically viable form.

For pre-enforcement planning, the first task would be statutory exposure. Counsel would need to know whether the client falls within the bill’s definition of covered AI activity, what entity would be ordered to transfer shares, which shareholders’ interests would be diluted or reassigned, and what deadline or triggering event would make the obligation imminent. Without those facts, the constitutional theory remains abstract.

The second task would be identifying the property interests with precision. A complaint should not simply say “equity” and move on. It should separate economic ownership, voting rights, director-election rights, veto rights, transfer restrictions, charter protections, and bylaw provisions. The more clearly the pleading shows a forced transfer of control rights, the harder it becomes for the government to characterize the statute as an ordinary revenue measure.

The third task would be remedy selection. A company facing an imminent transfer obligation would likely seek declaratory and injunctive relief before shares are conveyed. The papers would need to explain why the injury is not adequately repairable after the fact: once voting control, board seats, and veto power move to the government, corporate decisions can be made under a governance structure the plaintiff says is unconstitutional.

The best Takings Clause challenge to Sanders’ AI ownership proposal is credible because it focuses on what the statute transfers. Voting stock, board participation, veto leverage, and displaced governance protections are not the usual incidents of a cash tax. They are property and control interests. But the claim is not a guaranteed winner. The Supreme Court has not decided this exact statutory form, and the government would have nonfrivolous arguments based on the taxing power, public-benefit structure, and the regulated nature of the targeted AI activity.

References

1. News: Sanders Introduces Legislation to Create $7 Trillion AI Sovereign Wealth Fund — Bernie Sanders Senate, June 18, 2026.
2. The Public Should Own Half of the Big A.I. Companies — Bernie Sanders Senate.
3. An End to AI Competition: Senator Sanders’ Plan — American Action Forum.
4. The Fatal Conceit Gets a GPU Cluster: Bernie Sanders’ Plan to Socialize AI — Truth on the Market, July 3, 2026.
5. Bernie Sanders’ Dangerous and Unconstitutional Plan to Expropriate AI Firms — Reason, June 5, 2026.