Doosan Enerbility recently announced a massive order to supply seven large-scale gas turbines for a U.S. data center project.
At the heart of this deal is the insatiable demand for electricity from the artificial intelligence (AI) boom. Data centers, the brains behind AI, are consuming power at a staggering rate. Projections from agencies like the IEA suggest that by 2030, data centers could account for nearly 17% of all electricity used in the U.S. This has created immense pressure on tech companies to secure their own stable, large-scale power sources, leading them to build power plants right next to their data centers, a strategy known as 'on-site generation'.
Several recent events created the perfect conditions for this order. First, a crucial market signal came from the 'PJM capacity auction'. Prices in this auction, which ensures future grid reliability, hit a record high, largely due to the anticipated demand from new data centers. This made relying on the public grid an increasingly expensive and uncertain option. Second, a powerful political signal emerged from a White House meeting where the president openly encouraged big tech companies to build their own power generation facilities. This provided political validation for large on-site gas power projects.
Perhaps the most decisive factor, however, was the global supply bottleneck for large gas turbines. Industry giants like GE Vernova and Siemens Energy have their production schedules, or 'slots', fully booked until 2028 or even 2029. Their 'lead times'—the period from order to delivery—have stretched to several years. This created a significant opening for Doosan Enerbility. By guaranteeing a delivery schedule starting in May 2029, with one turbine per month, Doosan offered a clear solution to customers facing long waits and uncertainty.
In conclusion, Doosan's major order wasn't a sudden event but the result of converging forces: soaring AI-driven power demand, clear market and political signals favoring self-generation, and a critical supply shortage from competitors. This positions Doosan as a vital player capable of bridging the power gap for the rapidly expanding digital infrastructure.
- On-site generation: The practice of producing electricity at the same location where it is consumed, rather than purchasing it from a distant power plant via the electrical grid.
- PJM capacity auction: A market mechanism used in a large region of the eastern U.S. to ensure there will be enough power generation resources available to meet peak electricity demand in the future. High prices signal a potential supply shortage.
- Lead time: The total time that elapses between a customer placing an order and the product being delivered. Long lead times for turbines indicate high demand and limited manufacturing capacity.