“Will the future of humanity be determined by a handful of billionaires who have promoted and developed AI, with virtually no democratic input, who stand to become even richer and more powerful than they are today?” — Senator Bernie Sanders, writing in the New York Times.
Senator Sanders’ sovereign wealth fund proposal
The senator proposes creating a U.S. sovereign wealth fund by taking “50% stock in AI companies such as Anthropic, OpenAI and xAI.” The stated rationale is twofold: public ownership would give the government “the power, through its voting shares and an equal representation on each company’s board, to block decisions that hurt our citizens and to push for policies that help them”; and it would return “trillions of dollars potentially generated by AI” to the public.
Why public ownership could backfire
The authors who reviewed the plan agree with Sanders’ goals — democratic influence over AI and sharing AI’s economic gains — but caution that public ownership risks entangling corporate profit and the public interest. They argue it would “incentivize the government to clear regulations, permit the exploitation of workers and users, suppress competition, encourage AI adoption regardless of the responsibleness of the implementation or appropriateness of the use case, and otherwise act on behalf of corporate interests.”
They give a concrete hypothetical: “if growing, say, Nvidia from its first $5tn in value to its next $5tn also represents a doubling in value of this segment of the sovereign wealth fund, then you can expect the fund managers to support chip sales, foreign and domestic, with the same zeal as the company’s private investors.” In short, fund managers’ fiduciary incentives may push governments toward corporate-friendly policies.
Historical parallels: Norway’s fund and public pensions
The essay points to real-world examples. Ownership stakes by “the Norwegian sovereign wealth fund, the world’s largest,” have not apparently redirected oil companies toward pro-environmental policies; instead, “the Norwegian government’s dependence on those companies has inhibited them from taking climate action.” The authors draw a parallel to U.S. public employee pension funds, where “the fiduciary duty to generate wealth overwhelms any intention to direct their corporate holdings in the public interest.”
Alternatives: taxation and an AI public option (and the Swiss example)
Rather than large equity seizures, the authors recommend separating the goals of public influence and public reward. For revenue, they point to taxation — Senator Elizabeth Warren’s proposed excise tax on datacenters’ energy use and proposals for an “AI token tax” are cited as direct ways to make AI companies pay for social disruption.
To reshape AI development, the authors propose an “AI Public Option”: governments creating publicly developed and operated AI models “run by public institutions under democratic control” that set a competitive baseline for private offerings, analogous to a healthcare public option. They highlight a Swiss example: Apertus, “a large language model built by Swiss public servants, researchers at Swiss universities, using appropriately licensed training data and pre-existing Swiss public supercomputing infrastructure powered by renewable energy.” Apertus does not match the latest OpenAI and Anthropic models on benchmarks, but it outperforms them in “transparency, sustainability and compliance with EU regulations including adherence to copyright.”
They caution against conflating public AI with marketing-driven “sovereign AI,” which they describe as often “invoked as a marketing scheme for big tech companies looking to sell to governments; it demands public investment without guaranteeing public control.”
Political economy: why AI billionaires might accept the plan
The authors also interrogate why AI owners might support an extraordinary expropriation. They note Sanders’ claim that “the Trump administration and the billionaire owners of AI are aligned to the idea,” and suggest a motive: the tech owners may calculate that “for every dollar ceded to government stock expropriation, they will get back more in favorable government policies to protect that newfound investment.” In other words, an exchange of equity for softer regulation or other advantages could be the real bargain being contemplated.
What this means for Anthropic, OpenAI, and xAI; for U.S. policymakers; and for Swiss public institutions
- Anthropic, OpenAI, and xAI — as named targets of the proposal — would face a structural shift: equity dilution and potential board seats for government appointees, but also the prospect that public ownership could align fund managers’ commercial incentives with government policy choices.
- U.S. policymakers — confronted with the twin objectives of public control and public revenue — have alternatives that do not rely on equity seizure: energy excise taxes, AI token taxes, and establishing a publicly run AI option to exert competitive and regulatory pressure.
- Swiss public institutions — exemplified by Apertus — demonstrate a practical model: develop public AI with licensed data and public compute, prioritize transparency and regulatory compliance, and accept that performance parity with private leaders may come later than gains in accountability and sustainability.
The authors close where they began: the concentration of wealth and power in AI is a grave democratic risk, but they argue that seizing corporate equity is a blunt instrument that could worsen the problem. Instead, they urge combining taxation with a public AI option to capture economic rents and shape AI development under democratic control. “We urge Sanders and other political leaders to consider them.”
