Data center electricity demand will rise 165% by 2030, according to Goldman Sachs research.1 That single projection is restructuring how institutional investors value US utility and grid infrastructure stocks.
The driver is AI compute. A single ChatGPT query consumes roughly 10 times the energy of a Google search.2 Training next-generation large language models requires power equivalent to small cities.2
Microsoft, Amazon, and Alphabet are among the largest consumers on the US electricity grid.2 Their data center energy demand is structurally price-inelastic — compute must run regardless of power costs. That dynamic is fundamentally different from residential or commercial electricity demand, which responds to price signals.
The US grid was engineered for 1-2% annual demand growth.2 That design assumption is now a binding constraint. Existing grid-connected generation capacity near major data center hubs carries scarcity value that utilities did not price into long-term planning.
Power purchase agreements (PPAs) with AI hyperscalers are emerging as the key earnings quality metric for utility investors. Utilities that executed long-term supply contracts before the AI demand surge — particularly those with generation assets near Northern Virginia, Phoenix, Chicago, and Dallas data center corridors — carry the strongest forward earnings visibility.
Bitzero Holdings (AIBZ) is among the companies targeting this infrastructure intersection.2 The investment thesis: proximity to AI workload demand creates a recurring, durable revenue base that traditional industrial or residential customers cannot replicate.
Capacity factor utilization is the operational signal to watch. Grid-connected generation near high-density compute clusters runs at higher utilization rates, compressing unit economics and improving asset returns over time.
The primary risk is regulatory lag. US utility rate structures were not designed for large industrial AI customers. Interconnection queues and cost-allocation frameworks are under review in multiple states as grid strain increases — adding permitting uncertainty to new capacity build-out timelines.
Goldman Sachs projects 165% data center power demand growth through 2030.1 That implies a structural shift in who the marginal electricity buyer is. Utilities near established AI infrastructure corridors, with contracted supply agreements in place, are the direct beneficiaries of that shift.
Sources:
1 Goldman Sachs Research, data center power demand outlook
2 AI energy infrastructure market analysis, June 2026
