Who Owns the AI Boom? As Trillions Accumulate, Pressure Mounts to Share the Wealth

On August 16, The New York Times published a piece whose headline could define the second act of the AI story: “With A.I. Riches at Stake, Pressures Mount to Share the Wealth.” The article surveys a rapidly consolidating political reality — policymakers around the United States are maneuvering to ensure that a handful of tech giants are not the only beneficiaries of what may be the largest wealth creation event in modern economic history.

The timing is not accidental. In June, SpaceX completed a $75 billion initial public offering at a $1.77 trillion valuation — the largest IPO ever — and saw its market capitalization exceed $2 trillion on the first day of trading. Anthropic’s investors are now openly targeting a $2 trillion public debut of their own as soon as October. OpenAI’s valuation has climbed to roughly $852 billion, with tender offers pushing ever higher. Elon Musk became the world’s first trillionaire in June, even as workers face higher prices and mounting fears of AI-driven job displacement. The juxtaposition is becoming politically unbearable — and Washington, along with state capitals, has noticed.

The $7 Trillion Question: Sanders’ AI Sovereign Wealth Fund

The most aggressive proposal on the table comes from Senator Bernie Sanders, who in June introduced the American AI Sovereign Wealth Fund Act. The bill’s mechanism is startling in its simplicity: a one-time, 50-percent tax on the largest AI companies in the United States — a tax not on their profits, but on something far more valuable: their equity, paid in company stock.

Under the plan, the federal government would receive stock in companies such as OpenAI, Anthropic, and xAI, placing those shares into a newly established sovereign wealth fund. Based on current valuations, that fund would be worth approximately $7 trillion. Sanders’ office estimates that a 5 percent annual dividend from the fund could provide a direct payment of more than $1,000 to every person in America — and if AI valuations keep compounding, that number grows accordingly.

The proposal deliberately echoes the Alaska Permanent Fund, which has distributed oil royalties to Alaska residents annually since 1982. The argument is the same, updated for a new resource: the American public provided the data, the infrastructure, the education systems, and the public research that made modern AI possible. In a June op-ed titled “A.I. Is a Public Resource. You Deserve Your Share,” Sanders argued that “the public should own half of the big A.I. companies” — or at least half of the value they create.

Critics, including the Cato Institute and the Tax Foundation, call the plan expropriation by another name — a confiscatory equity tax that would chill investment, invite capital flight, and face near-certain constitutional challenges. Even sympathetic observers note the governance thicket: a federal fund holding large minority stakes in private companies creates conflicts of interest that make existing tech-government entanglements look quaint.

OpenAI’s Counterproposal: A 5 Percent Stake for Washington

The industry is not waiting to be taxed. According to a July report by the Financial Times, later confirmed by CNBC and Reuters, OpenAI has discussed voluntarily handing the U.S. government a 5 percent equity stake in the company. At the $852 billion valuation from its recent tender round, that holding would be worth roughly $42.6 billion.

The strategic logic is transparent. OpenAI faces intense political scrutiny in Washington — over its data-center deals, its energy consumption, its restructuring, and its competitive conduct. A government stake would align federal incentives with the company’s success, converting a regulator into a shareholder. As one person familiar with the talks put it, the proposal is designed to “clear political obstacles” — the same instinct that led banks to accept Treasury ownership during the 2008 crisis, though in this case the company is offering rather than being forced.

The contrast between the two approaches is instructive. Sanders’ plan would take half of the AI industry for the public. OpenAI’s would give away a twentieth of one company on its own terms. Both, however, accept the same premise the Times article identifies: purely private ownership of frontier AI is becoming politically untenable.

The Data-Center Backlash Fueling the Moment

What gives these proposals traction is the physical spread of AI infrastructure — and the backlash trailing it. The Times notes that more than 700 data centers are under construction across some 40 states, with Amazon, Google, Microsoft, Meta, and Oracle projected to spend about $750 billion this year on data centers, chips, and related AI infrastructure — a figure expected to reach $1.5 trillion across this year and next.

Communities from Michigan to Virginia report surging electricity prices, strain on water supplies, and local tax abatements that often leave residents subsidizing the very facilities disrupting them. New York Governor Kathy Hochul signed a one-year statewide moratorium on the largest new data centers in July — the first of its kind. A separate Times investigation documented air and water pollution impacts on host communities. President Trump, for his part, has shrugged off the concerns, staying “especially bullish” on data centers even as voter anxiety grows, and his administration has moved to repurpose federal land for AI facilities and their power plants.

This is the political soil in which wealth-sharing proposals germinate. When the costs of AI — energy, water, land, labor displacement — are local and immediate, while the gains are global and concentrated, the demand for redistribution becomes structural, not cyclical.

From UBI to “Universal Basic Capital”

The policy menu is widening beyond one-off funds. California Governor Gavin Newsom has floated giving workers a stake in AI gains — a “universal basic capital” spin on universal basic income. The Atlantic examined similar models in July. Senator Elizabeth Warren has proposed taxing the energy usage of AI data centers and investing the proceeds in people. Brookings published a public-finance framework arguing that sovereign wealth funds invested in AI companies “capture returns that automatically scale with AI’s economic impact” — elegantly hedging the uncertainty: if AI is overhyped, the funds cost little; if it transforms the economy, the public shares automatically.

The intellectual through-line is a shift from income redistribution (UBI checks funded by taxes) to ownership redistribution (public equity stakes in the engines of growth). The Times piece suggests this shift is now mainstream enough that the question is no longer whether the public gets a share of AI wealth, but how large, through what mechanism, and on whose terms.

What Comes Next

Several inflection points will decide whether August 2026 is remembered as the moment wealth-sharing went from think-piece to statute:

• Anthropic’s October IPO. A $2 trillion public listing would crystallize paper gains into institutional fortunes — and hand wealth-sharing advocates their most vivid exhibit yet.
• The fate of the OpenAI stake talks. A voluntary 5 percent transfer, even informal, would set a precedent every frontier lab would have to answer for.
• State-level experimentation. With Washington gridlocked, states are becoming laboratories — data-center taxes in Congress, accountability mandates in Washington State and Oregon, dividend schemes patterned on Alaska’s.
• The 2026 midterms. Candidates in both parties are discovering that “who owns AI” polls better than anyone expected.

The deeper tension the Times identifies may be irresolvable within current market structures. Frontier AI is being built on public data, public research, public infrastructure, and publicly subsidized energy — but its ownership is almost entirely private, and its gains are compounding faster than any technology in history. Sanders’ $7 trillion fund may never pass. OpenAI’s 5 percent concession may never close. But the direction of travel is set: the question of who owns the AI boom has entered the political mainstream, and it is not leaving.

For an industry that has grown accustomed to moving fast and answering to no one, that shift may prove as consequential as any model release this year.