By LaiLa Kearney and Liz Hampton
New York/Houston (Reuters)-Varales of US power, utility and natural gas companies away Monday in some of the largest one-day drops recorded, as new AI technology from Chinese initial Deepseek initiated a projected surge Calls US electricity and technology expenditure.
Power producers were among the largest winners in the S&P 500 last year on the expectations of balloon demand from the energy -legged data centers needed to the scale of Big Tech's Big Tech artificial intelligence technologies.
Adopting AI models more widely developed by Deepseek, which it says is built in less than two months and is cheaper than models currently used by US companies, Generally demanding electricity and leading to reduced power construction, analysts and economists say.
“If proven to be proven, the efficiency used within the Deepseek open source model can be applied by the hyperscalers to their models, which would lead to a more modest demand,” analysts with Evercore ISI said in a note.
Large technology companies, also known as the developers of hyperscaling data centers, have devoted tens of billions of dollars in the development of AI data centers over the past year.
In the US, data centers consumed about 4.4% of electricity in 2023 but are predicted to use 6.7% to 12% of all power by 2028, according to a report produced by the Lawrence Berkeley National Laboratory.
The independent power provider Constellation Energy sank, whose shares shot up about 100% in 2024 mainly on its ability to sell nuclear power and gas to US data centers, about 20% in trading on Monday after news about news Deepseeek developments.
Vistra was down 30% and an Energy Corp talented competitor was down 22%.
Deepseek AI could also threaten the dominance of current AI leaders, located in Silicon Valley, and slow down their use of data centers. AI Deepseeek help overtook the US competitive ChatGpt in downloads from the Apple App store on Monday.
But with the wider adoption of AI, even with more energy -efficient models, the demand for power could surprise everywhere, said Ed Hirs, an energy economist at the University of Houston. He warned that sales of power stocks could be blind and short -lived.
“In this case, if Deepseek turns out to be what everyone wants, and they sell to US companies, and US companies are changing their algorithms to adopt it, ' n mean a bigger, faster wider development, “Hirs said.
Still, electricity companies, and even feed producers involved in power generation, were under pressure.