Global stock markets suffered a sharp decline three days after the US midterm elections, with technology companies bearing the brunt of the sell-off as investors reassessed the prospects of the artificial intelligence industry. Companies listed on the technology-heavy Nasdaq Composite collectively lost approximately $2.5 trillion in market value, while the S&P 500 fell by more than 4 per cent. The last time the S&P 500 experienced a decline of comparable magnitude was in April 2025, when US President Donald Trump announced his tariff policy. Markets in South Korea and Japan also recorded substantial overnight losses, while Germany’s Dax fell at the start of Friday trading, despite having reached a record high earlier in the week.
The immediate trigger was growing concern that the political consequences of the Republican Party’s defeat could slow the expansion of the American AI industry. Joachim Klement, a capital markets strategist at Panmure Liberum, described the sell-off as a reaction to fears that a strengthened Democratic Party would impose restrictions on the sector. According to the technology investor Philipp Klöckner, 77 per cent of American voters would prefer to see future data-centre construction restricted. “The average American sees AI as a risk. And data centres are the physical manifestation of AI,” he said.
The prospect of tighter restrictions has raised concerns about the construction of facilities needed to support the industry’s expansion. Klöckner estimates that decisions may have to be reconsidered over whether and where approximately 50 per cent of planned data centres in the United States can be built. Even where projects eventually proceed, delays are likely. Such uncertainty has affected construction companies and hardware suppliers, which fear that expected profits from the expansion may not materialise on schedule. Shares in Nvidia fell by more than 9 per cent, while Intel lost approximately 10 per cent, contributing to the wider decline in the Nasdaq and S&P 500.
The losses extended beyond the largest chipmakers to companies directly involved in the infrastructure supporting artificial intelligence. Semiconductor manufacturer Micron, storage producer Sandisk and Elon Musk’s space and AI company SpaceX were among those affected. SpaceX is also planning to manufacture its own chips, making it exposed to developments in the sector’s investment cycle.
The political uncertainty is particularly significant because Democrats are expected to control both chambers of Congress from 2027. Although the party is not fundamentally opposed to artificial intelligence, it favours stronger regulation than the relatively permissive approach adopted under Trump. However, this does not necessarily mean that data-centre construction will come to a halt. Carsten Brzeski, chief economist at ING, said he could not imagine Democrats wanting to cripple what is currently the largest growth engine of the American economy.
The more immediate financial danger, according to Klöckner, concerns the relationship between chip purchases and the construction of the facilities required to use them. Companies have already ordered substantial quantities of hardware, but if data-centre projects are delayed, the equipment may remain unused. Operators must nevertheless continue paying interest on the loans used to finance their purchases. This could create difficulties for heavily indebted businesses such as Nebius and CoreWeave, although Klöckner believes major technology companies such as Amazon and Google are not exposed to the same level of danger. Even if smaller operators fail, he expects their businesses to be acquired by larger companies, limiting the overall impact on demand for chips.
For the wider market, the current correction could represent a normalisation of expectations rather than the collapse of the AI industry. Klöckner argues that a decline of between 5 and 10 per cent in share prices does not constitute a breakdown of the sector. He believes a gradual reduction in excessive valuations could actually prevent the AI bubble from bursting abruptly. Brzeski also maintains that artificial intelligence will remain “the most disruptive technology of the coming years”, while acknowledging that not all the industry’s ambitious profit expectations can be realised. Futures prices ahead of the opening of US markets were already indicating a modest recovery among some of the technology shares most severely affected by the sell-off.
Klement, however, has a considerably more pessimistic outlook. In an October forecast for 2027, the capital markets strategist predicted that global markets could lose approximately 30 per cent of their value, with technology-heavy indices such as the Nasdaq potentially falling by half. He believes signs of weakening momentum in the AI boom have been visible since the spring. Shares in major technology companies, including Google, Microsoft and Amazon, have come under pressure when they announced particularly large investments in artificial intelligence. “The market is sending a signal: we like that you are investing in AI, but please do not invest too much,” he said.
The increasing reliance on borrowing to finance data-centre expansion is another source of concern. Financing costs are rising, making large investment programmes more expensive. Klement believes a further interest-rate increase by the US Federal Reserve this year could become another catalyst for the end of the current boom.
The international consequences were visible in Germany, where the Dax had climbed to 26,692 points on Thursday, establishing a new record following the US election results. By Friday morning, however, it had opened approximately 2 per cent lower as the technology sell-off spread across markets. Felix Rohse, an asset manager from Hamburg, said the initial positive reaction and the subsequent decline were not contradictory. Investors had initially welcomed the election outcome, but their attention had since shifted towards its potential economic consequences. He argued that concerns about stricter regulation and restrictions on new AI data centres were prompting investors to reassess risks that had already existed. The Dax’s decline also demonstrated the close integration of international equity markets.
The market turbulence has also affected expectations surrounding Anthropic, the developer of the Claude AI model, which was preparing for a stock market flotation in November. The company had been targeting a valuation of approximately $1.5 trillion. Klement had hoped that its planned listing would provide another boost to the AI investment boom, but said recent market developments suggested that investor enthusiasm was weakening. According to the Wall Street Journal, three major investors who had considered participating in the flotation no longer regarded the proposed valuation as realistic. One reportedly placed the company’s potential value closer to $1 trillion.
Klöckner nevertheless believes that proceeding with the flotation could be preferable to postponing it, since the company’s financial performance might not improve and delaying the listing could eventually make it more difficult to complete. He also suggested that slower data-centre construction could have an unexpected commercial advantage for Anthropic. Reduced infrastructure spending could postpone the repayment of some loans while limiting the overall availability of computing capacity. Scarcer computing resources could, in turn, allow AI services to be sold at higher prices.
The sell-off also carries implications for investors holding exchange-traded funds tracking major international indices. The MSCI World, a widely used benchmark for global equity investments, derives almost three-quarters of its exposure from US stock markets. Nvidia alone accounts for approximately 5 per cent of the index. Emerging-market funds are also exposed to the AI industry through major semiconductor manufacturers, including Taiwan’s TSMC and South Korea’s Samsung and SK Hynix. A sharp decline in a heavily weighted sector can therefore affect portfolios far beyond the United States.
Nevertheless, major indices gradually adjust their composition as the relative market values of companies change. If a sector loses importance over time, companies from other industries can acquire greater representation. Rohse said this mechanism limits investors’ dependence on a small number of individual businesses, although it does not protect them from immediate market losses.
Financial advisers quoted in the analysis recommend that investors avoid making hurried decisions in response to individual days of falling share prices. Rohse argues that people investing over the long term, with an allocation to equities appropriate to their financial circumstances and tolerance for risk, should not react precipitously. Investors whose portfolios are heavily concentrated in individual AI companies, however, may need to reconsider their exposure.
Klement follows a specific indicator when deciding whether to reduce risk in his own portfolio: if the S&P 500 falls below its 200-day moving average, he reduces his equity holdings. On Thursday, the index came close to crossing that threshold. When he sells shares, he reduces exposure across his portfolio rather than targeting technology companies alone, moving the proceeds into government bonds, including German federal bonds. For investors wishing to remain in equities, he identifies pharmaceuticals, healthcare, food, tobacco and beverage manufacturers as sectors less dependent on artificial intelligence, alongside companies benefiting from government spending on infrastructure and defence.
