Sharp falls in shares of some major chip makers have raised investor concerns that the recent surge in enthusiasm around artificial intelligence-related companies may be easing, according to BBC News.
The rally has been driven by expectations that AI could reshape work and daily life, while creating large gains for investors in a small group of companies, mainly in the United States and Asia. However, recent declines in businesses that supply key AI infrastructure have led some investors to reassess the scale and timing of potential returns.
BBC News reported that some of the steepest falls have been in Asia. Shares in South Korean chip makers SK Hynix and Samsung have fallen by 46% and 35% respectively over the past month, amid concerns that recent demand for AI chips may not be sustainable. Even after those falls, the shares remain threefold and fivefold higher respectively over the past year.
Concerns have also affected large US technology companies. Google and Tesla shares briefly dropped before recovering last week after both companies signalled plans to spend billions more on AI, despite the technology not yet generating profits for them.
Investors are also watching results from Meta, Microsoft and Amazon for further evidence of how much major companies are committing to AI. On Wednesday, the Nasdaq closed around 9% below its June record high, with worries about heavy AI spending among the factors weighing on the index.
Russ Mould, investment director at AJ Bell, told the BBC there remained “a healthy degree of scepticism” about whether AI-related investment would deliver matching returns.
Technology investor Eileen Burbidge said the market had not yet reached a severe turning point. “The AI bubble hasn’t burst but it’s letting out air,” she told the BBC.
BBC News also reported that investor caution has been sharpened by a reported manufacturing breakthrough by a Chinese company, which could make China more self-sufficient in chip design and production. Wider concerns remain over whether leading AI companies will be able to charge users enough to justify the hundreds of billions being spent on chips and data centres.
There are further questions over the costs of maintaining AI infrastructure, as data centres are expected to need regular upgrades to use newer and faster processors. Some governments are also pausing, restricting or banning new data centre construction on environmental grounds linked to water and energy use.
Despite the recent volatility, Burbidge said investors who bought chip maker shares a year ago would still be in a strong position. The market, however, is now scrutinising AI spending plans more closely than during the peak of investor excitement.
