Power availability is a key constraint to AI growth. Computers need to be turned on. That idea has sent data center operators scrambling for internally controlled power-generation solutions, helping send shares of power-equipment providers to the moon.
Shortages can create opportunities for secondary market players. That appears to be happening in power. These days, a natural gas power generation turbine might be tougher to score than Taylor Swift or Beyoncé concert tickets.
"Scalpers emerge in the turbine queue," wrote GLJ Research analyst Gordon Johnson on Wednesday.
He noted that a power turbine delivery slot had been sold on a secondary market. For a buyer who wants power sooner, the turbine will cost almost 50% more than a theoretical list price, according to Johnson, who cited a Semianalysis report.
The turbine maker wasn't cited. Large gas turbines are manufactured by the likes of GE Vernova and Siemens Energy. GE Vernova didn't immediately respond to a request for comment about transferring slots. Any slot transfer would likely require a turbine maker's approval.
The sale highlights just how tight the market for power generation has become. Turbine makers are essentially sold out until 2030. Some 75 gigawatts of power generation equipment are on order. That's a lot. Those assets could generate about 660 terawatt hours of electricity annually, or about 15% of current U.S. demand.
Many of those turbines have been ordered by AI data center operators, and will be "behind the meter," meaning they aren't hooked up to the grid, with operators paying a utility.
The emergence of a secondary market is bullish, demonstrating sky high expectations for AI-related power demand, admits Johnson. But it doesn't remove the risk of canceled orders if the AI or economic cycle turns lower.
Hyperscalers are spending hundreds of billions a year building AI computing capacity. A downturn doesn't look imminent, but political regulation and public backlash-from higher utility and water bills, could lead to an AI slowdown.
Johnson rates GE Vernova stock Sell and has a $470 price target for shares. It's the lowest target on Wall Street by a wide margin, according to FactSet.
Overall, almost 80% of analysts covering GE Vernova rate its shares Buy. The average Buy-rating ratio for S&P 500 stocks typically ranges from 55% to 60%. The average analyst price target for GE Vernova stock is about $1,240.
GE Vernova stock was up 1.8% in early trading Wednesday at $897.87, while the S&P 500 was up 0.2%. Gains left shares up 45% over the past 12 months, but down 15% over the past month.
Investors have been a little nervous about the AI trade lately, with GE Vernova shares paying a price.