OpenAI’s Sam Altman, US President Donald Trump, US senator Bernie Sanders, and California governor Gavin Newsom do not agree on much, yet they have all proposed that the United States create a sovereign wealth fund ( SWF ), seeded by stocks from AI companies, to cope with AI’s negative impacts. It is a good idea. Unfortunately, their proposals are woefully inadequate.
Consider Altman’s plan, which would have the largest AI companies – Anthropic, Google, Meta, Microsoft, Nvidia and OpenAI – give the US government a 5% stake. At current valuations, this would translate into holdings worth around US$800 billion, which would be used as seed money for the creation of an SWF.
It is an innovative idea, but hardly a transformative sum. Altman’s SWF is designed to be a one-time “wasting trust”: the Treasury gets US$800 billion worth of stock up front, but at some point the fund is expended.
To be sure, the US would not have to sell all those shares right away, instantly depleting the fund. As with a trust fund, it could instead sell one-tenth of the shares each year, and let the remainder grow in value. Conservatively, that would produce about US$1.4 trillion over 10 years, assuming the stock value increases by 10% per year, the US stock market’s historical average.
But US$1.4 trillion would not come anywhere close to bridging the US government’s financing gap, at a time when the national debt is on track to exceed US$48 trillion by 2030. Merely maintaining existing Social Security benefits will require an additional US$2.6 trillion by 2032. If this gap is not bridged, beneficiaries will face a 22% haircut.
Given this, the 5% stake Altman advocates is nowhere near adequate. Even the Sanders plan, which would be funded by a 50% tax on AI firms’ stock, would not generate nearly enough revenue to meet America’s mounting financing needs.
Other global SWFs are significantly larger than those envisioned by Altman and others, especially relative to the size of the relevant economy. Energy-rich Norway, whose population is less than 2% the size of America’s, boasts the world’s largest SWF, with US$2.2 trillion in assets. China has two SWFs totalling about US$3.5 trillion. Kuwait, Saudi Arabia, Singapore and the United Arab Emirates have SWFs worth about a trillion dollars each.
Moreover, other countries’ SWFs are self-replenishing “permanent funds”. That makes a big difference. Norway’s enormous investment-based portfolio accrues such large returns that it contributes more to Norway’s public budget than oil profits do. This is the kind of fund the US should design.
The first step toward creating a federal permanent fund ( FPF ) would be the US government’s establishment of the trust that would oversee it. The government would then raise the seed money by selling Treasury notes to the public on a month-to-month basis. By selling US$200 billion worth of Treasuries each month for 10 years, the trust could accrue US$24 trillion, which it could use to purchase 20% stakes in each of the three major stock index funds—the S&P 500, the Nasdaq, and the Dow—such as through an Exchange-Traded Fund.
At the aforementioned 10% annual compound return rate, the value of these investments would increase by almost 90% over 10 years, to US$45 trillion. With the historical market yield on 10-year Treasury notes around 4.5% ( though it hovers near 5.2% today ), the FPF would earn a net compound return of around 5.5% per year over the 10-year incubation period. After that, the FPF would sell the index fund shares month after month and pay off the outstanding US$24 trillion principal, leaving the FPF with around US$21 trillion.
At that decade mark, the self-replicating FPF would be extended indefinitely, raising about US$2.1 trillion every year. The result would be the world’s largest SWF, capable of addressing America’s long-term funding needs, including paying down the federal debt, funding the Social Security gap, and covering AI-related unemployment benefits. This fund could even help pay for other pressing needs, such as affordable housing, guaranteed basic income, health care and more – all without raising taxes or increasing government debt.
This “universal capitalism” plan would require the US to sell significantly, but not exorbitantly, more Treasury notes annually. In 2025, the US sold US$3.4 trillion in Treasuries; another US$2.4 trillion per year would amount to a 70% increase. The sales would happen gradually, on a month-to-month basis. And since the notes would be secured by ownership of appreciating stock investments, which will eventually be worth more than the Treasuries, the government’s net debt would not increase.
Investing in index funds is key, as it would both prevent the FPF from interfering with individual companies and ensure that one firm’s fall from grace could not drag down the fund’s value. The S&P 500 has proven to be more stable, and has earned higher returns over the last 10 years, than the stocks of former high-flying companies like AIG, AT&T, Citigroup, General Electric, Hewlett-Packard, IBM and Pfizer. The size of the FPF’s index purchases should be of little concern, since the “big three” asset managers ( BlackRock, Vanguard and State Street ) already collectively hold more than 20% of the S&P index and are the largest shareholders in over 90% of all S&P 500 companies.
The one downside of this approach is that the FPF’s investment returns would not be available during the 10-year incubation period. Here is where the much smaller Altman plan could be useful: it could provide temporary benefits for the first decade, until the FPF is fully loaded. Ultimately, however, that plan can provide little more than a stopgap. A much larger, permanent SWF remains essential to enable the US government to capitalize on private-sector dynamism to cover financing shortfalls and meet Americans’ needs at a time of rapid economic and technological change.
Steven Hill is a fellow at Rutgers University’s Institute for the Study of Employee Ownership and Profit Sharing and John Menke is the president of the Menke Group.
Copyright: Project Syndicate
