This piece pushes back on a dangerous idea: turning private AI companies into public property. It argues that handing pieces of these firms to the government risks saddling taxpayers with massive losses while locking the nation into an unsustainable, power-hungry data center model.
Americans should be skeptical when Silicon Valley shows up at the taxpayer trough asking for permanence. The story here is simple: a private, loss-making sector wants public ownership so its risks become someone else’s problem. That is a recipe for moral hazard and permanent government programs that grow and never shrink.
Reports say OpenAI has discussed a plan where tech firms would contribute equity to a sovereign-wealth-style fund so the public would “share in the sector’s soaring valuations.” That sounds noble on the surface, but the math and incentives tell a different story. When losses, liabilities, and ongoing subsidies are bundled with public ownership, the taxpayer ends up guaranteeing what private capital should be forced to price.
The president reportedly floated the idea that “pieces” of AI companies could be “given to the American public” to soothe concerns over rapid deployment. He even said “the public will become very rich.” Those words deserve scrutiny, not cheerleading. Promises of windfalls rarely survive the spreadsheet, especially when the industry itself says it needs a government safety net.
OpenAI’s executives have been candid about needing a government “backstop,” and the company asked for “grants, cost-sharing agreements, loans, or loan guarantees” to scale. Asking for taxpayer-funded guarantees under the banner of competing with China is clever politics, but it is not sound economics. The state cannot fix a business model that depends on never-ending capital infusions and resource-intensive infrastructure.
Generative AI demands absurd amounts of compute, power, and space, and that expense doesn’t vanish if the public holds equity. Tokens and GPU time cost real money, and firms are already seeing those costs explode. When firms push employees into paid AI platforms, the token bill can outpace salary costs, and even big companies have ordered engineers to stop using tools because the cost-to-utility ratio fails basic common-sense accounting.
There is a better technical path that the current build-out ignores: localized edge computing with low latency and far less dependence on hyperscale data centers. Instead the industry pours concrete, grabs land, and burns capital on facilities that may not leave durable public value. Unlike railroads or broadband, many of these installations will depreciate quickly and provide little long-term public benefit.
Investors and developers are racing to list these companies publicly and shove them into indexes so trillions in pension money flows their way. That converts private risk into systemic risk. Once pensions and government programs are tied to a fragile, resource-hungry industry, every bailout, subsidy, and regulation will be framed as protecting ordinary Americans, when in truth it’s protecting insiders.
Some of these players are even seeking special favors like federal land to lower costs and speed build-outs. When capital structures depend on taxpayer concessions and loan guarantees, the project is not a market success story, it is a campaign to socialize losses. We learned similar lessons from green-energy mandates: if the only way to profit is through government favors, the product should face harsher scrutiny, not more subsidies.
There is a clear conservative case against turning AI into a public program. Let markets sort winners and losers, insist on transparency in how taxpayer funds are proposed to be used, and protect pensioners from speculative bets. Public skepticism, not equity in risky startups, is the proper default when private models depend on public rescue.
If AI firms are viable, they should build profitable businesses without asking taxpayers to underwrite land, power, or losses. If they are not viable, the right answer is not to make them permanent government projects. Handing over ownership or guarantees swaps innovation for dependency, and that will be the real legacy if this plan proceeds.
Americans should demand a different path: technological leadership rooted in competition, accountability, and prudent stewardship of public resources. The stakes are too high to accept a quiet deal that turns venture risk into permanent taxpayer liability.
