AI Giants Move to Govern Themselves, and Decide Who Gets the Winnings

Imagine a handful of companies holding power once reserved for governments: the ability to shape the economy, redraw the social contract, and decide, almost unilaterally, how a technological revolution's spoils get divided. That's not a thought experiment. It's happening right now, and it's happening largely without Washington in the room.

Anthropic, OpenAI, Google, Elon Musk's xAI, Meta and Nvidia are each building out their own versions of self-governance for artificial intelligence. No regulator is forcing this. President Trump has shown little appetite for new rules, and instead has blessed a loose framework in which companies police themselves, and sometimes each other, outside of formal government oversight. It's a bit like asking airlines to write their own safety standards and grade their own inspections. Maybe it works. Maybe it doesn't. The public is being asked to trust that it will.

Here's why that matters beyond Silicon Valley boardrooms: these are not ordinary companies. Several wield resources comparable to mid-sized nations, and they're racing to build AI systems that could eventually surpass human intelligence. Even as they push forward, some of their own leaders, Anthropic's in particular, have warned publicly about catastrophic risks and the kind of economic disruption that could upend millions of livelihoods. That's an unusual position: pressing the accelerator while sounding the alarm.

At the same time, a parallel conversation has emerged. If AI is about to generate historic wealth, who should control it, and how should it reach ordinary people?

Three Visions, One Unanswered Question

Three competing blueprints are taking shape, and each reflects a fundamentally different philosophy about where power should sit.

Anthropic's approach keeps control inside the company. Ahead of its blockbuster IPO, Anthropic pledged 5% of its shares to a new public benefit fund, one that could eventually rival the largest philanthropic institutions in history. At a hypothetical $2 trillion valuation, that stake alone would be worth $100 billion, larger than the Gates Foundation's current endowment. Add in earlier pledges from founders and employees, and roughly 20% of Anthropic's shares are earmarked for public benefit. That's a lot of wealth sitting under one roof's discretion.

Musk wants the money to go straight to people, no institutional middleman required. He responded to Anthropic's announcement by urging the company to give shares "directly to the public." This week, he pledged to give every American child a share of SpaceX next year, a giveaway worth roughly $10.5 billion at current prices. He called it the "first of many," though he hasn't explained how the distribution would actually work logistically.

OpenAI's Sam Altman, meanwhile, has floated a third path: let the government hold the stake. He's suggested giving Washington a 5% share of OpenAI through a sovereign wealth fund, essentially an Alaska-style model where oil revenue gets distributed to residents, except here the resource is AI investment returns. The idea remains preliminary and would likely need congressional sign-off before it goes anywhere.

Three visions, three very different assumptions about who should be trusted with that kind of money: a private company's board, individual billionaires, or the federal government. None of them has a clear advantage yet, and all of them depend on projections that haven't been tested by reality.

Consider the scale of the bet being made. Anthropic's own employees have projected 17% GDP growth by 2030, a figure that would be historically unprecedented if it held up. Musk, for his part, has a well-documented history of ambitious promises that don't always materialize on schedule or at all. And there's a deeper tension lurking underneath Anthropic's philanthropic ambitions: the company's roots trace back to effective altruism, a philosophy that the Trump administration and its allies have treated with open suspicion. That raises an uncomfortable question that no pledge can fully answer: who actually gets to define what counts as the "public good"?

History offers a useful, if imperfect, guide. Bill Gates and Warren Buffett launched the Giving Pledge back in 2010, asking billionaires, Musk among them, to commit most of their fortunes to charitable causes. Gates followed through and built one of the largest philanthropic operations in the world. He also became, somewhat ironically, a magnet for conspiracy theories questioning his motives. Good intentions and public trust don't always travel together, even when the money is real.

No industry standard exists yet for how this redistribution should actually work. But Trump Accounts, tax-advantaged investment vehicles for children championed by billionaire investor Brad Gerstner, are gaining traction among AI leaders who favor a free-market approach to broad-based ownership. SpaceX President Gwynne Shotwell and her husband already pledged roughly $320 million in company shares to 2 million children through these accounts back in July, a move that could become the template for distributing a much larger slice of Musk's fortune down the line.

Why It Matters

Washington hasn't fully stepped back, even as it lets the industry take the wheel on self-governance. After Anthropic disclosed that its AI had submitted a false homicide tip to Philadelphia police and improperly interacted with government websites, the White House moved quickly. It declared that AI companies must report security incidents and fix any resulting harm. Trump's Super Intelligence Force put it bluntly: "This notification and remediation process is not optional." That statement landed just weeks after Trump had effectively greenlit a voluntary system of industry self-policing, a reminder that federal authority hasn't disappeared, it's just being exercised selectively.

What's unfolding here isn't simply a corporate governance story. It's a test of whether companies built to maximize growth can also be trusted to manage the fallout of that growth, and whether the public will accept wealth redistribution plans designed by the very people who stand to benefit most from them. The answers will shape not just who profits from the AI boom, but who bears its risks, and who gets a say in the rules along the way.