Anthropic announced plans to commit 5 percent of its shares to a new Public Benefit Fund, with more potentially added as the company’s value grows. At the roughly $2 trillion valuation reportedly being targeted for its IPO, that initial stake would be worth about $100 billion, making it one of the largest charitable commitments in history. The company says the fund could even make direct payments to specific individuals, among the other priorities.

Anthropic has put real equity behind a promise the AI industry has made from the start: that this technology should benefit everyone, not just those who build it. The announcement also comes amid a growing fight over who should share in AI’s wealth. Senator Bernie Sanders has proposed a 50 percent tax on shares in major frontier AI companies, with the proceeds directed to a federal wealth fund. OpenAI has floated giving Washington a stake of its own. Anthropic’s voluntary commitment offers a better way forward.

Start by giving America’s children a stake in the AI economy.

The infrastructure to do this already exists. Trump Accounts, created in the tax law President Trump signed last year, provide each eligible child born from 2025 through 2028 with a $1,000 federal deposit invested in a low-cost stock index fund. Families, employers and philanthropists can contribute more, giving children nearly two decades of compound growth before they reach adulthood.

Two recent changes from the Treasury Department make these accounts especially promising. First, Treasury automatically established accounts for more than 60 million children. Previously. only seven million had enrolled, including just 5 percent of low-income families. When Michael and Susan Dell pledged $6.25 billion to contribute $250 into the accounts of 25 million children, only 1.2 million had received it. Now their gift reaches all of them, thanks to automatic enrollment. The contribution finds the child, not the other way around.

Treasury also opened a pathway for qualifying charities to contribute publicly traded shares directly into these accounts. Once Anthropic goes public, its equity could help fund contributions through eligible charitable organizations, subject to the applicable rules. The companies would not necessarily have to liquidate all their pledged shares up front.

Anthropic could dedicate one-fifth of its new fund’s initial equity stake commitment, equivalent to 1 percent of the company, to these children’s investment accounts. OpenAI should make a comparable commitment.

Together, they could begin by matching the federal government’s $1,000 contribution, giving eligible newborns a $2,000 starting investment. They could then provide additional contributions to children in the lowest-income communities, where families have the fewest resources to invest. And when federal contributions end after the 2028 birth cohort, the companies could keep seeing accounts for children born after that. It would cost billions, but less than the price of a single gigawatt-scale data center, and represents only a fraction of the wealth these companies expect to create.

When I served in the George W. Bush White House, we talked about building an ownership society. The idea was that people who own a piece of the economy have a stake in its success. That distinction matters in this debate. A federally managed wealth fund would concentrate ownership and influence in Washington, giving politicians a new lever over companies they regulate. A forced transfer of half a company’s equity could undermine the private investment financing America’s lead in AI and slow the race against China. Children’s investment accounts would put assets directly in families’ hands instead.

This idea has deep bipartisan roots. Democrat Senator Bob Kerrey proposed KidSave Accounts in the 1990s with backing from both the Heritage Foundation and the Progressive Policy Institute. Republican Rick Santorum and Democrat Chuck Schumer later teamed to propose newborn investment accounts two decades ago. Cory Booker championed baby bonds, while Ted Cruz helped write Trump Accounts into law. Across decades and political parties, the principle that has endured is giving children the opportunity to begin life with an asset that can grow over time.

None of this is a substitute for responsible AI development and deployment. Public trust cannot be bought with investment accounts, and the labs still have an obligation to prove their systems are safe, reliable, and broadly beneficial. But a generation with a financial stake in AI-driven prosperity will have a reason to want it to succeed.

The AI community has a word for making powerful systems serve the interests of people: alignment. They have spent billions trying to solve it inside their models. Their IPOs offer a chance to practice it outside them.

Investors are about to make the biggest financial bet in history: that AI will build the future. Tech companies should make a bet on the children who will inherit it. Give kids a stake.