Few would have expected the rare political agreement between Bernie Sanders and Donald Trump. Bernie Sanders is a U.S. senator from Vermont, known as a longtime advocate of social justice and a critic of economic inequality. The two men, otherwise at opposite ends of the political spectrum, found themselves on a similar wavelength in June 2026: the U.S. government should acquire stakes in the largest companies developing artificial intelligence. And the proceeds should go directly to American citizens.
Sanders’s plan: 50 percent in shares
Senator Bernie Sanders of Vermont published an extensive article announcing that he would soon introduce a bill called the American AI Sovereign Wealth Fund Act. The largest U.S. AI companies, specifically OpenAI, Anthropic, and xAI, would have to pay the government a one-time tax of 50 percent. Not of their profits, but of their shares. These would go into a government sovereign wealth fund that would manage the returns for the benefit of all Americans. The result would be direct payments to citizens. And, over time, according to Sanders, funding for healthcare, education, and housing as well.
Sanders built his argument around one fundamental idea: artificial intelligence did not simply emerge out of nowhere. Models such as ChatGPT are built on books, texts, newspaper articles, research papers, works of art, and the conversations of millions of people. Mostly without permission and without compensation. This was, after all, confirmed by OpenAI CEO Sam Altman himself, who said that the models had been trained on “the collective experiences and knowledge of humanity.” “The creative work of millions of people has essentially been stolen by some of the wealthiest people in the world. It is time to make things right,” Sanders wrote.
Trump is listening
What might once have seemed like a left-wing fantasy is now also under scrutiny by the White House. NOTUS was the first to report that senior Trump administration officials had held preliminary talks with major AI companies about a possible government stake. Altman reportedly brought the idea directly to Trump as early as the beginning of 2025. When reporters asked the president about it aboard Air Force One, Trump replied: “There is something very interesting about it, where it becomes almost a partnership with the American people. We’ll look at it.”
The Trump administration has already taken similar steps. Last year, the United States acquired a 10 percent stake in Intel. Since then, Intel’s shares have risen fivefold, and the original $10 billion investment is now worth more than $50 billion. The government has also acquired similar stakes in several quantum computing companies, including IBM.
Altman and Sanders even met in person. Altman reportedly agreed with the idea of public ownership in AI, but considered 50 percent too high a bar. Nevertheless, he indicated a willingness to collaborate on the principle.
The idea also has serious opponents
Harvard economist Nathan E. Sanders and security technologist Bruce Schneier acknowledged in a commentary for the British newspaper The Guardian that they share Sanders’s goal. At the same time, however, they proposed a different approach.
According to them, government ownership of shares in AI companies creates a dangerous conflict of interest. If the value of a sovereign wealth fund depends on the growth of companies such as Nvidia or OpenAI, the government will naturally have an interest in protecting their expansion rather than overseeing it. As an example, they cited Norway’s sovereign wealth fund, the world’s largest at more than $2 trillion, which owns large stakes in oil companies. In their view, these ties have not prevented Norway from remaining dependent on fossil fuels—quite the opposite.
Their alternative? Taxing the energy consumption of data centers, as Senator Elizabeth Warren proposes, or so-called public AI—models operated by the government under democratic oversight. They cite Switzerland as an example, where the Apertus model was created by public institutions using licensed data and powered by renewable energy.
Harvard economist Ken Rogoff, meanwhile, warns of a situation in which governments that fail to secure a place in the AI supply chain could face mass unemployment without the tax revenues or capacity to manage the situation. “No one really knows what such a world would look like, let alone how to hold it together,” he wrote.
Waves of nationalization and AI as public property
Professor Nicholas Mulder of Cornell University argues that since 2020, we have been going through the fourth wave of nationalization in the past hundred years. Since then, governments around the world have taken over assets worth more than half a trillion dollars, the first such increase since the 1970s. At the same time, governments increasingly view AI and quantum computing technologies as strategic national resources, much like oil or energy. And where there is a strategic interest, the question of control follows.
In this light, Sanders’s plan may seem less exotic. After all, OpenAI itself endorsed the idea that a public fund should allow citizens to share in the returns from AI in its April 2026 industrial policy proposal. Anthropic proposed government sovereign wealth funds backed by stakes in AI. Elon Musk wrote that direct government payments were the best solution to unemployment caused by automation.
