New York: Rebounding tech stocks drove US indexes higher on Tuesday, January 28, a day after they tumbled on doubts about whether the artificial-intelligence frenzy really needs all the dollars being poured into it.
The S&P 500 climbed 0.9% to claw back more than half of its earlier drop. The Dow Jones Industrial Average added 136 points, or 0.3%, and the Nasdaq composite rallied 2% after sliding 3.1 % the day before.
The spotlight remained on Nvidia, whose chips are powering much of the move into AI and whose stock has become a symbol of the surrounding frenzy. It rose 8.8% after plunging nearly 17% the day before, which was its worst drop since the 2020 Covid crash.
Other AI-related companies also held steadier, including chip company Broadcom, which rose 2.6%. Constellation Energy picked up 1.4% after plummeting nearly 21% on Monday. It had earlier rallied on expectations it will help supply the electricity that vast AI data centres would gobble up.
Such revenues are threatened after DeepSeek, a Chinese company, said it was able to develop a large language model that can perform as well as big
US rivals but at a fraction of the cost.
That raises questions about whether all the spending expected for AI chips and electricity will need to happen.
AI-related stocks have been Wall Street’s biggest stars in recent years, soaring on expectations that big spending will only continue to grow. The gains, though, also created criticism that the stock prices had simply gone too high, too fast.
It’s still uncertain how much DeepSeek’s development will upend the AI industry. While it could mean less growth in spending than expected for data centres, electricity and chips, it could also boost other areas.
“If AI becomes less expensive to use, we think businesses will adopt it more quickly, making a greater investment in AI software,” according to James Egelhof, chief US economist at BNP Paribas.
“We think this acceleration in adoption could mean a rise in software investment that offsets – or even dwarfs – any deceleration in spending on data centre structures, hardware and related investment, he said.”