The AI boom is shifting from chips to kilowatts. U.S. data-center electricity use could more than double to about 376 terawatt-hours by 2030 from roughly 167 TWh in 2023, according to a Yahoo Finance analysis of government and industry data. That surge is turning power from a background expense into a gating factor for growth, pushing battery storage into the core of AI infrastructure. Developers plan to add 24 gigawatts of utility-scale batteries in 2026—second only to solar—underscoring storage’s role as a time-shifting buffer rather than a source of new generation. Investors are stretching the AI trade beyond semiconductors to servers, software and storage, with deals such as Ford’s battery arrangement drawing attention. Storage won’t solve the supply gap, but it can smooth volatility and improve reliability on a grid built for steadier loads.
The Power Grid Is AI’s Next Bottleneck
- Trending
- Comments
- Latest
