AI Cannot Scale Without Solving the Power Problem First
AI runs on electricity, and there is not enough of it. Goldman Sachs projects that AI data center power consumption will rise 175% by 2030. Sightline Climate found that up to 50% of announced data center projects face delays, with power access being the primary bottleneck. Of 190 gigawatts of planned data center capacity, only 5 gigawatts are under construction. The gap between what AI companies need and what the grid can supply is one of the biggest constraints on AI progress. Grid-scale batteries are emerging as a critical piece of the solution.
The Scale of the AI Power Problem
- AI data center power demand expected to rise 175% by 2030
- Only 5 GW of 190 GW planned data center capacity is under construction
- 36% of data center projects saw timelines slip in 2025
- Gas turbine shortages and an aging grid limit new power generation
- The U.S. will have nearly 65 GW of battery storage by the end of 2026
How Batteries Help Bridge the Gap
Grid-scale batteries store energy when supply exceeds demand and release it when demand peaks. For data centers that need consistent, uninterrupted power, batteries provide a buffer against grid variability. Form Energy’s 100-hour battery, which stores energy using iron-air chemistry, is one of the most ambitious projects in this space.
Google signed a 1.9 GW clean energy deal in Minnesota that blends wind and solar with a 30 GWh battery from Form Energy. This hybrid approach reduces dependence on the grid while maintaining the reliability data centers require.
Google’s deal included working with utility Xcel Energy on a new rate structure that encourages the adoption of new energy technologies in the utility’s planning process, setting a precedent for other partnerships.
On-Site Power Is Growing
Some of the largest data centers are bypassing the grid entirely. Less than a quarter of projects that have identified a power source use on-site or hybrid power, but together they represent 44% of total planned capacity. That means the biggest projects are leading the shift toward energy independence.
Amazon, Google, and Oracle are building data centers with on-site solar, wind, and battery storage. Nuclear power, including small modular reactors, is attracting renewed interest for providing reliable, carbon-free baseload power.
The Investment Opportunity in AI Energy
Energy infrastructure companies serving the AI sector face high demand, long-term contracts, and physical barriers to entry that software companies do not have. A startup that can reliably deliver power to data centers has a more defensible business model than most AI application companies.
Form Energy is raising $500 million ahead of an IPO. Startups building solid-state transformers, grid management software, and advanced battery chemistries are attracting venture capital. For investors looking at the AI ecosystem, the power supply chain offers a way to profit from AI growth regardless of which model or company wins the market.
AI’s future depends on electricity. The companies solving the power problem will be as important to AI’s success as the companies building the models.
