AI Data Centers: Catalysts of Rising Global Energy Demand

Energy-intensive artificial intelligence (AI) data centers are on track to sharply increase global power consumption, with estimates suggesting a 160% hike in data center power usage by 2030, according to a Goldman Sachs report. This marks a transformative shift in global energy dynamics following a period of stagnant demand.

The report highlights that merely generating power is not the sole challenge; transmission bottlenecks pose significant hurdles to bringing new power plants onstream. In the United States, for instance, natural gas powers most data centers due to its availability, yet challenges in permitting and supply chains often delay new plant operations by 5 to 7 years.

Meeting the projected data center demand growth requires new capacity, distributed across various sources: 30% from natural gas combined cycle gas turbines, another 30% from natural gas peakers, 27.5% from solar energy, and 12.5% from wind. Although natural gas is a key power source, renewables are expected to play an increasing role, offering quicker and more efficient solutions than traditional gas-based plants.

Tech giants are navigating these energy challenges by opting for a hybrid power approach, strategically using multiple sources, and cautiously investing in long-term solutions like nuclear energy. Notably, companies are opting for strategies like forward-start power purchase agreements to expedite progress, as exemplified by Alphabet's agreement with Elementl Power to position three sites for advanced nuclear energy development.