Trump and Sanders Reach Unusual Agreement: The Government Should Own a Piece of Anthropic

Washington rarely finds common cause between its two poles, but AI ownership has managed it. The Trump White House spent much of late May negotiating a deal under which OpenAI, xAI, and other frontier AI companies would donate equity to a new U.S. sovereign wealth fund. Two weeks later, Senator Bernie Sanders introduced legislation that would impose a 50 percent tax on equity stakes in OpenAI, Anthropic, and xAI and funnel the proceeds to a public wealth fund. The mechanism differs, the political logic differs, and the targets partially overlap. What does not differ is the destination: both proposals end with the federal government holding a financial stake in the companies building the most consequential technology in a generation.

Anthropic sits at the center of one proposal and at the edge of the other, a position that reveals something important about how the company has navigated Washington over the past two years.

Two Paths to the Same Address

The White House approach was brokered in part through the "Stargate" consortium framework and emerged from a broader negotiation over whether AI companies could sidestep federal oversight if they contributed to national AI infrastructure. Under the reported terms, OpenAI would donate equity to the sovereign wealth fund voluntarily as part of its conversion from a nonprofit hybrid to a for-profit company. The arrangement was framed publicly as a patriotic contribution, a way for American AI companies to ensure that the gains from AI accrued to citizens rather than solely to shareholders and foreign investors.

Anthropic was not part of those discussions. Sources familiar with the negotiations say the company's ongoing legal dispute with the Pentagon, which placed Anthropic on a restricted vendor list over export-control compliance questions, made including it in a White House-adjacent deal politically awkward. A spokesperson for Anthropic declined to comment on whether the company was invited to participate.

Why Anthropic Isn't at the Table

The exclusion from White House negotiations maps closely onto Anthropic's complicated federal relationship. The company was placed on a Pentagon restricted vendor list earlier this year while simultaneously holding active contracts with the NSA and other intelligence agencies. The contradiction reflects a federal bureaucracy that has not yet developed a coherent policy toward AI vendors: different agencies within the same government arrived at opposite conclusions about whether Anthropic could be trusted with sensitive work.

That ambiguity has made Anthropic a difficult partner for politicians who need clean narratives. OpenAI, whatever its governance controversies, has not been formally blacklisted by any federal department. xAI benefits from Elon Musk's personal relationship with the administration. Anthropic has neither of those anchors, and the scrutiny over its Middle East investor relationships has added a further complication.

"These companies should not be allowed to privatize the gains while the public absorbs all the risks. If we are going to fund this technology with taxpayer infrastructure and public research, the public deserves a share of what they helped build." Senator Bernie Sanders, statement accompanying the Public Wealth Fund Act, June 2026

The Anthropic Paradox

There is a certain irony in Anthropic's position. The company was founded on the premise that AI development done responsibly, with safety at its center, would produce better outcomes for everyone. That safety-first posture attracted capital, built credibility in academic and policy circles, and helped push Anthropic's valuation to approximately $965 billion ahead of a potential IPO. But the same stance that generated commercial value has also made Anthropic the most regulated target in the room. Sanders named it specifically in his bill because it is a recognizable symbol of concentrated AI wealth. The Pentagon put it on a restricted list because its safety research produced documentation that compliance reviewers found concerning.

The company's deliberate distance from Washington may have been rational strategy in 2022, when the AI industry was still building its D.C. lobbying infrastructure. It looks more costly now that equity deals are being negotiated at the White House. Companies at that table will shape the terms of the sovereign fund arrangement, including which categories of equity count as a contribution, what restrictions attach to government ownership, and how fund governance is structured. Anthropic will inherit whatever framework those companies agree to, without having influenced it.

The Sanders bill, by contrast, is not something Anthropic can negotiate around. The legislation names the company by valuation tier, meaning that any company crossing a threshold that Anthropic already exceeds would be subject to the 50 percent equity tax. Whether the bill passes in anything close to its current form is a separate question; as a statement of political intent, it signals that the left flank of the Democratic Party has identified Anthropic as a target, not a partner.

A federal court has already issued a preliminary ruling suggesting that at least one element of the broader AI equity framework, a provision requiring companies to register equity transfers with a new federal AI oversight body, likely violates the First Amendment as applied to private speech about corporate structure. That ruling adds legal uncertainty to an already unstable policy environment, and Anthropic's legal team will be navigating it without the goodwill that comes from having participated in the original negotiations.

The company's most plausible path forward runs through its IPO timeline. A public offering would transform the equity question from a political negotiating chip into a matter of securities law, giving Anthropic the shareholder structure and disclosure obligations of any other public company. Whether that transition happens before or after the sovereign fund framework solidifies will go a long way toward determining how much of Anthropic's future the federal government claims a right to define.