Former Google CEO Eric Schmidt is warning that the U.S. may need an additional 92 gigawatts of electricityto keep up with AI growth. His point is simple: AI isn't just a software boom anymore - it's a power and infrastructure challenge.
Data centers run 24/7, require heavy cooling, and are scaling fast. Recent projections cited in coverage suggest U.S. data centers used 4.4% of electricity in 2024 and could reach 12% by 2028, with higher power costs likely in data-center-heavy markets.
Some tech firms are already responding by pledging to cover their data center electricity costs, while others are exploring longer-shot ideas like space-based data centers powered by solar.
Why it matters for buyers
The pace of AI adoption is now colliding with the slower reality of grid buildout. The next phase of AI growth will depend on where power can be secured, how fast infrastructure can be added, and how operators procure it. For data center, AI compute, and adjacent industrial buyers, the implication is direct: dedicated behind-the-meter generation is moving from a niche option to a mainstream procurement path.
Smartland Energy continues tracking how digital demand is reshaping energy markets, grid planning, and behind-the-meter strategy. See how we structure dedicated power for data center buyers →