Senator Bernie Sanders just dropped the most ambitious tech policy proposal in American political history: legislation that would create an AI sovereign wealth fund giving the public a direct 50% ownership stake in major AI companies through a one-time stock tax.
Whether you think this is visionary or disastrous, it forces a conversation the tech industry has been trying to avoid: if AI was trained on humanity's collective knowledge, who should profit from it?
Sanders' proposal would apply to the largest AI firms—think OpenAI, Anthropic, Google DeepMind, Meta AI. These companies would be required to transfer stock to a public fund, giving American taxpayers direct ownership and dividend rights. The mechanism would be a one-time tax on stock holdings above a certain threshold.
The senator's argument is straightforward: AI models are trained on public data—books, articles, code, images, conversations—that represent humanity's collective intellectual output. Private companies have monetized this public resource without compensating its creators. A sovereign wealth fund would ensure the public shares in the returns.
The tech industry's response is predictable: this would destroy American AI competitiveness, drive companies overseas, and kill innovation. Sam Altman has already suggested as much in past comments about aggressive AI regulation.
But here's what makes this proposal interesting: several other countries already have sovereign wealth funds that own stakes in strategic industries. Norway's Government Pension Fund owns stakes in thousands of companies globally. Singapore's Temasek and GIC manage hundreds of billions in strategic investments. The model isn't radical—applying it to AI is.
The economics are genuinely complicated. AI companies argue they invested billions in compute, research, and talent to transform raw data into useful products. That's true. But they also benefited from decades of publicly-funded research, open-source infrastructure, and yes, freely available training data that they never paid for.
There's also the small matter of whether these companies will actually generate returns worth owning. Despite billions in investment, most AI companies still don't have sustainable business models. is reportedly burning cash at an alarming rate. is venture-funded and loss-making. The public might be buying equity in the world's most expensive science experiments.
