EDITORIAL: Wall Street’s sudden discovery that the artificial-intelligence revolution needs voters’ permission as well as investors’ money has exposed an uncomfortable weakness in its favourite trade. The Democrats’ sweeping US midterm victories, accompanied by widespread opposition to data-centre development, helped turn previous weeks Wednesday’s relief rally into Thursday’s rout.
The Nasdaq 100 lost roughly USD 2.5 trillion in market value in a single session before technology shares recovered some ground on Friday. For a boom priced on seemingly limitless expansion, the prospect of political and financial limits was enough to cause panic.
It would not be rash to declare that the AI bubble has burst, or is even bursting. The technology’s economic potential remains questionable, and a market correction offers no definitive verdict on future earnings. But the week’s convulsions have supplied fresh ammunition to those questioning the extraordinary valuations and even more extraordinary spending commitments behind the boom. Voters concerned about electricity bills, water consumption and enormous computing facilities have added a political obstacle to an investment story already struggling to explain how its returns will justify its costs.
The spectacle recalls Carmen Reinhart and Kenneth Rogoff’s 2009 book, This Time Is Different: Eight Centuries of Financial Folly, which examined how successive generations persuaded themselves that familiar financial excesses had somehow become harmless. Their warning resonates whenever soaring asset prices are defended by the claim that a new era has rewritten economic arithmetic. The irony this time is that even investors from the heart of the technology industry are questioning the exuberance. Jack Selby, managing director of Thiel Capital — the family office of Peter Thiel, PayPal and Palantir co-founder, Facebook’s first outside investor and one of Silicon Valley’s most influential financiers — has warned that the AI bubble could burst before year-end, citing extravagant capital requirements, debt financing and fanciful fundraising narratives.
The more consequential question concerns the creditors financing that exuberance. Data-centre construction is increasingly dependent on bonds, bank lending, private credit and off-balance-sheet vehicles. The Bank of England has warned that this expanding web of borrowing, sometimes involving opaque and circular arrangements, could amplify losses if AI earnings disappoint. Its September assessment cited an estimate of around USD 450 billion in global AI-related debt issuance by early September, more than double the whole of 2025. Banks are exposed directly as lenders and indirectly through their relationships with investment funds and other financiers.
A collapse in the expected revenues supporting these projects could therefore travel well beyond technology shareholders. Falling collateral values, refinancing problems and defaults would test institutions that lent against assumptions of relentless AI growth. The IMF’s latest financial stability work has warned of vulnerable banks and deeper links to non-bank lenders, although it also finds the banking system broadly resilient under its stress scenarios. A systemic crisis is a possibility to examine rather than a foregone conclusion. Precisely because the damage cannot yet be measured, regulators should be establishing who ultimately bears the risk.
And then comes the political irony. OpenAI’s finance chief discussed the possibility of a government backstop for infrastructure financing last year, later clarifying that the company was seeking no such guarantee. The controversy nevertheless demonstrated how quickly an industry celebrating private innovation can entertain public support for its financial architecture. With Democrats poised to control Congress and a Republican president still in the White House, any rescue involving taxpayers could become a bruising political contest just when speed and clarity matter most.
Should an AI bust seriously damage major lenders, would Washington rediscover the doctrine that certain institutions are too big to fail? Would investors who enjoyed the upside be protected from the downside, leaving ordinary taxpayers to foot the bill? The 2008 crisis taught the world what happens when financial institutions become indispensable after accumulating risks nobody adequately scrutinised. AI may well transform the economy. The disturbing possibility is that the financial system surrounding it could require a distinctly old-fashioned bailout.
Copyright Business Recorder, 2026
Comments