You may already be bailing out the AI business
Washington is treating the industry as if it’s too big to fail, even as the market sends lukewarm signals.
Is an artificial-intelligence bubble about to pop? The question of whether we’re in for a replay of the 2008 housing collapse—complete with bailouts at taxpayers’ expense—has saturated the news cycle. For every day that passes without disaster, AI companies can more persuasively insist that no such market correction is coming. But the federal government is already bailing out the AI industry with regulatory changes and public funds that will protect companies in the event of a private sector pullback.
OpenAI’s chief financial officer, Sarah Friar, said the quiet part out loud at a Wall Street Journal event last week when she told her interviewer that the company is looking to governments to “backstop" loans for AI chip purchases with a “guarantee" that will elicit private financing. It would be bad public policy not to, in Ms. Friar’s accounting. By presenting U.S. leadership in AI as a moon shot on which American national security and economic growth will depend, Ms. Friar breezed past the core worry that has investors balking: Is building AI at such a large scale worth the financial risk?
This is a live issue as AI companies assert a need to build out vastly more chips, data centers and other infrastructure. It’s unclear if the demand for AI can cover the cost. Consultants at Bain & Co. in September estimated that cloud service providers—firms like Google, Microsoft and Amazon that have powered the AI boom—will have to generate an additional $2 trillion in annual revenue by 2030 to afford all the necessary infrastructure. That is more than five times as large as the global market for software subscriptions. In 2024 Amazon, Alphabet, Apple, Meta, Microsoft and Nvidia combined made less than $2 trillion.
A little taxpayer-backed financial security could sound quite good to AI-related companies. Ms. Friar was light on the details last week, and after the Journal event she backtracked, saying her use of the word “backstop" had “muddied" things. OpenAI chief Sam Altman tweeted that the company doesn’t want a bailout. But it’s notable that in October an open letter by OpenAI’s chief global affairs officer, Christopher Lehane, called for federal loan guarantees for AI-related infrastructure.
The pressure to perform is becoming unprecedented for these firms, and the evidence suggests they won’t get a passing grade. An MIT study this summer found that of about 300 organizations that bought or built their own generative AI tools, 95% reported zero return on their investments. AI adoption has declined at large companies, according to Census Bureau surveys. Some observers have also pointed to what looks like circular deal making—a hallmark of the telecom bust. Nvidia is investing $100 billion in OpenAI, which plans to buy millions of Nvidia chips. That sort of funding loop could get dangerous if the market cools.
Despite the lukewarm market signals, the U.S. government seems intent on backstopping American AI no matter what. It’s a stance that predated this administration. The 2024 AI National Security Memorandum recast the success of U.S. AI as the national security priority of our times—making AI companies officially too big to fail.
The Trump administration is rolling out the red carpet for these firms. The administration’s AI Action Plan aims to accelerate AI adoption within the government and military by pushing changes to regulatory and procurement processes. Government contracting offers stable, often lucrative long-term contracts—exactly what these firms will need if the private market for AI dips. This spring, the Atlantic reported that the Department of Government Efficiency was attempting to automate civil-service jobs with AI. On top of that, the One Big Beautiful Bill Act authorized about $1 billion in AI funding, and the administration says more loans, grants and tax incentives for AI infrastructure will shortly follow.
Washington is trying to shield AI companies from regulations. The White House’s Action Plan seeks to limit federal and—through funding reviews—state regulations on AI. The administration is looking to fast-track regulatory processes for AI infrastructure and plans to open thousands of acres of federal land to data centers. The administration has even threatened tariffs against the European Union for digital regulations. And AI companies seem to want more: Mr. Altman has raised the idea of “AI privilege" to give chat histories a confidentiality akin to conversations with a lawyer or doctor.
The U.S. AI bailout doesn’t stop at the border. The Trump administration last month kick-started the American AI Exports program through which the White House will mobilize tech, financial and diplomatic resources to push U.S.-origin AI onto new foreign markets—specifically selling AI to foreign governments. In the U.K., despite feeble noises about the need for national champions, London signed a deal to use OpenAI models this summer.
The AI industry is leaning in. In May, OpenAI announced its OpenAI for Countries initiative, which looks to expand other nations’ data-center capacities and ChatGPT adoption. Nvidia CEO Jensen Huang has been going around the world giving presentations to government leaders about the need for their countries to have their own “sovereign AI," which will certainly require many chip purchases.
There are many risks of giving AI special treatment. One is that government authorities may dole out favors, choosing winners and losers. Deepening ties between AI leaders and the state could further concentrate power within the tech industry among current star companies.
Another is a market correction. Federal policy has jumped the gun: We don’t yet know if AI will transform the economy or even be profitable. Yet Washington is insulating the industry from all sorts of risk. If a bubble does pop, we’ll all be left holding the bag.
Ms. West and Ms. Kak are executive directors of the AI Now Institute.
topics
