Written by Laila Kirney and Lez Hampton
New York/Houston (Reuters)-US companies, utilities and natural gas that were sold on Monday in some of the largest drops of one day, as the new AI technology from the Chinese Deepseek technology raises doubts about an expected increase in the United States for electrical order and technical spending.
Energy producers were among the largest winners of the S&P 500 last year due to amplification expectations from databases wandering in the energy needed to expand the scope of artificial intelligence technologies in Big Tech.
The broader adoption of artificial intelligence models such as those developed by Deepseek, which it says was created in less than two months and cheaper than the models currently used by American companies, may lead to a lesser extent to demand electricity in general and lead to a smaller energy building, analysts and economists said.
“If it is proven correct, the competencies used in an open source Deepseek model can be applied by their strait to their models, which will lead to a more overseeing request,” said analysts at Evercore ISI in a note.
Large technology companies, also known as the developers of the super data center, have developed tens of billions of dollars in developing the artificial intelligence data center during the past year.
In the United States, databases consumed approximately 4.4 % of electricity in 2023, but 6.7 % to 12 % of energy are expected to be used by 2028, according to a report submitted by Lawrence Berkeley National Laboratory.
The independent Constellation Energy, whose shares increased by 100 % in 2024 to a large extent to its ability to sell nuclear energy and gas to American databases, and sank about 20 % in trading on Monday after news of Deepseek.
Vestra decreased by 30 % and the Taleen Energy Corp competition decreased 22 %.
Deepseek ai can threaten the dominance of current artificial intelligence leaders, which is based in Silicon Valley, and slows down their data centers. Deepseek’s help from Apple Chatgpt has exceeded the competition from Apple App Store on Monday.
Ed Herz, an economic expert at Houston University, said, but with broader artificial intelligence, even with more energy -saving models, energy demand can rise everywhere. He warned that selling salad shares could be short and short -term.
“In this case, if it turns out that Dibsic is what everyone wants, sell it to American companies, and the American companies change their algorithms to adopt them, this means only a greater and wider development.”
However, electricity companies, and even the producers of raw materials related to power generation, were under pressure.
https://media.zenfs.com/en/reuters-finance.com/65b705f09d7788fa2a8c49379afafd91
Source link