As artificial intelligence companies prepare for massive public offerings, a novel proposal is gaining traction: instead of scattering shares to individual citizens or creating new public funds, route them into the existing Social Security system to help close its long-term funding gap.
Anthropic, which is reportedly eyeing a valuation between $1.5 trillion and $2 trillion, has pledged 5% of its shares to a Public Benefit Fund, with additional contributions tied to stock performance. Elon Musk has promised to give one SpaceX share to every American child. But a new opinion piece argues that the most efficient and equitable destination for such equity is the Social Security trust funds, which currently cannot hold corporate stock.
Bipartisan interest in public investment funds
The idea is not new. OpenAI has proposed a public wealth fund, and politicians from President Trump to Sen. Bernie Sanders have floated similar concepts. However, none have produced a working fund. Social Security, by contrast, is already operational, covering nearly every American worker and holding $2.6 trillion in reserves.
But the program is underfunded. The Trustees project the retirement trust fund will run dry by late 2032, triggering a 22% benefit cut. The 75-year shortfall stands at $29 trillion in present value. While AI equity pledges alone won't close that gap—Anthropic's 5% is worth roughly $75–100 billion—they could be a catalyst for larger contributions.
Why not borrow?
Sens. Bill Cassidy (R-La.) and Tim Kaine (D-Va.) have proposed borrowing $1.5 trillion to seed an investment fund. But borrowing creates financing risks, especially with long-term Treasury yields at their highest since 2002. Instead, the authors suggest contributing assets outright, eliminating those risks.
They point to existing federal assets: Treasury's stake in Fannie Mae and Freddie Mac, estimated at $400 billion, and the government's gold holdings, worth about $1.1 trillion. Together, that's roughly $1.5 trillion—enough to close a third to nearly half of the shortfall.
A model from philanthropy
The authors draw a parallel to Andrew Carnegie, who funded public libraries only if towns provided land and operating funds. Similarly, AI companies could condition their contributions on Congress creating a dedicated fund. Congress could build such a vehicle as part of a solvency deal or standalone legislation.
Washington already allows donated stock in federal programs—last year's Trump Accounts for children accept corporate contributions. If a child's account can hold stock, so should the program that serves 71 million Americans today.
The authors argue that Social Security was the institutional answer to a great economic dislocation, and the AI boom may be another such moment. By giving the public an account to receive equity, the first contributing companies could do for this century what Carnegie did for libraries.
