US officials have talked about acquiring government equity in AI companies.

      The concept is so unconventional that it takes a moment to comprehend: the United States government potentially acquiring equity in companies involved in developing frontier AI. A report from NOTUS reveals that senior officials in the US have engaged in initial discussions with leading AI companies regarding the federal government purchasing shares in these firms, which it is also attempting to regulate.

      The discussions include OpenAI. Sam Altman, the company's CEO, has been advocating in Washington for Congress to allocate funding for AI testing instead of mandating model approvals. He has periodically discussed the idea of government stakes with senior officials from the Trump administration since the president's second term began, initially presenting it to Donald Trump in early 2025 and revisiting it recently.

      According to the report, these conversations have focused on companies willingly giving shares to the government rather than the government directly purchasing them. Part of the pitch involves what the government would do with the proceeds. Options discussed include allocating the profits for public purposes, such as distributing a dividend to all American households, thereby transforming government ownership of AI into a means of distributing the technology's benefits to citizens.

      This idea is politically charged, especially since polls cited by NOTUS indicate that 55% of Americans believe AI will cause more harm than good in their daily lives. Additionally, OpenAI's corporate structure is currently scrutinized amid the Musk v. Altman trial regarding its transition from a nonprofit to a profit model.

      This public skepticism influences how companies are responding. Tech firms are seeking ways to gain the trust of Americans who are apprehensive about AI, and offering the public a stake, whether literal or financial, in these profits addresses concerns that benefits are being concentrated among a small group. From this perspective, voluntarily transferring shares can be seen as a measure of reputation management as well as fiscal policy.

      However, the main concern is structural and immediate. A government that possesses shares in a company it regulates serves as both a shareholder and a referee, which critics flagged as a conflict of interest. The motivation to safeguard the value of its holdings complicates the obligation to oversee the company, and no amount of framing around public dividends can entirely alleviate this tension.

      Moreover, this notion has emerged across the political spectrum. Proposals for public equity in AI have also been suggested from the left, arguing that if technology displaces many jobs, the public should have a claim on the resulting profits rather than allowing them to be monopolized by private shareholders. OpenAI itself has acknowledged this concern, even as Altman downplays the potential for an AI-driven job crisis.

      The fact that discussions between Trump administration officials and progressive advocates have converged on a similar idea—that the state could hold a stake on behalf of the public—illustrates the growing uncertainty about who truly benefits from AI.

      At this point, these are reports of discussions, not a formal policy or agreement. The talks are characterized as preliminary, the share transfers are described as hypothetical and voluntary, and the dividend is just one of several proposed ideas.

      However, the mere existence of these conversations highlights how much the relationship between the US government and the AI industry has evolved, moving away from distant regulation towards a potential future where the state may become a co-owner of the very sector it is supposed to regulate.