Chevron, GE Vernova Double Down on Powering Big AI Data Centers -- Barrons.com

Dow Jones
01-28

Avi Salzman and Mackenzie Tatananni

Chevron and GE Vernova will build enormous natural-gas power plants for artificial-intelligence data centers around the country that are big enough to power entire cities, the companies announced on Tuesday.

Chevron and gas-turbine maker GE Vernova are partnering with Engine No. 1, an investment firm that first came to prominence when it won a proxy fight to change the board of Exxon Mobil. They've been working on the project for more than a year, and already have put money down to secure turbines, said Chris James, founder and chief investment officer of Engine No. 1, in an interview.

The first "power foundries" will use seven GE Vernova turbines and could be up and running by 2027, the companies said. They did not release financial details about the partnership. GE Vernova stock was up 0.7% in early trading, while Chevron stock was down 0.1%.

The plants will be designed to be plugged directly into the data centers, so they won't need to connect to the larger electric grid, and siphon power away from existing consumers. That also means they won't need to wait for years for transmission hookups. President Donald Trump has already discussed these kinds of "behind the meter" configurations as a solution to the power needs of AI data centers. Trump's "energy emergency" executive order could smooth the way to these plants being approved.

"We're in very close contact with Trump and his team, who are very supportive of what we're doing here," said Jeff Gustavson, president of Chevron New Energies, in an interview.

Chevron, one of the country's largest producers of natural gas, expects to provide much of the gas for the plants. The companies may work with other producers depending on where the plants are located, however, Gustavson said. The companies involved in this effort are looking at five sites around the country, but they did not identify exactly where those sites are located. They have been in discussions with tech companies, but have not announced any specific partnerships with data-center owners.

Tech companies have said they want to become greener, and produce less carbon emissions even as they use more power for data centers. Last year, several tech names were talking about getting power from nuclear plants, because they don't emit carbon. Joining with big natural-gas plants could make it harder to achieve those environmental goals, though Chevron said that it will design the plants with the potential for them to capture the carbon they emit, and store it underground.

The announcement comes one day after stocks associated with AI -- including stocks of power companies that send electricity to data centers -- fell dramatically because of competition from China's DeepSeek AI program. DeepSeek appears to be able to run sophisticated AI models without as much electricity, which could dent the investment case for big power companies. GE Vernova stock fell 22% on Monday.

James of Engine No. 1 said that the DeepSeek news only strengthens his resolve that the U.S. will need to invest more in power, not less.

"I think this is this should be thought of as a wake-up call," he said. "We need to start building stuff here in the U.S. if we're going to win this AI battle."

Write to Avi Salzman at avi.salzman@barrons.com and Mackenzie Tatananni at mackenzie.tatananni@barrons.com

This content was created by Barron's, which is operated by Dow Jones & Co. Barron's is published independently from Dow Jones Newswires and The Wall Street Journal.

 

(END) Dow Jones Newswires

January 28, 2025 10:49 ET (15:49 GMT)

Copyright (c) 2025 Dow Jones & Company, Inc.

免責聲明:投資有風險,本文並非投資建議,以上內容不應被視為任何金融產品的購買或出售要約、建議或邀請,作者或其他用戶的任何相關討論、評論或帖子也不應被視為此類內容。本文僅供一般參考,不考慮您的個人投資目標、財務狀況或需求。TTM對信息的準確性和完整性不承擔任何責任或保證,投資者應自行研究並在投資前尋求專業建議。

熱議股票

  1. 1
     
     
     
     
  2. 2
     
     
     
     
  3. 3
     
     
     
     
  4. 4
     
     
     
     
  5. 5
     
     
     
     
  6. 6
     
     
     
     
  7. 7
     
     
     
     
  8. 8
     
     
     
     
  9. 9
     
     
     
     
  10. 10