Demand is only part of the problem
Berkeley Lab estimates that U.S. data centers used 192 terawatt-hours of electricity in 2024. Its June 2026 report projects 649 terawatt-hours in 2030 in the reference case, within a wider scenario range of 521–843 terawatt-hours. Those figures describe a substantial change in demand. They are projections, not a schedule of approved connections.
For a facility operator, the practical question is more specific: how much dependable power will be available at this location, and when? A national electricity forecast cannot answer that question. Neither can the nameplate rating of a generator or a battery.
Capacity, connection, and continuity
A useful power plan must connect three things. Capacity is the amount of power the facility needs. Connection determines what the grid can supply at the site. Continuity concerns what happens during interruptions, constraints, and changes in operating conditions. A weakness in any one can limit the value of the others.
This is why power availability belongs in the product specification for critical computing infrastructure. Operators need to understand the conditions under which a system can deliver power, how it recovers after an event, and what replenishment or maintenance it requires. Installed capacity alone leaves those questions unanswered.
Flexibility can change the question
Duke's 2025 Rethinking Load Growth study modeled 98 gigawatts of additional load within existing system peaks if 0.5% of that added load's potential annual energy were curtailed. This is a system-level estimate with important limitations, including network and generator operating constraints. It is not a promise that a particular data center can connect sooner.
The useful idea is that the timing of demand matters alongside its total size. A facility able to change its grid draw may have different options from one that requires the same uninterrupted supply in every condition. Whether that flexibility is practical depends on the actual site, operating agreement, and critical workload.
The question behind QEYS Reserve
QEYS Energy is developing QEYS Reserve, an approach that pairs rapid battery response with a replenishable on-site energy reserve. Our question is whether that combination can give a critical facility more control over its grid demand while preserving the work it needs to perform.
The answer must come from engineering evidence and customer requirements. Demand growth creates urgency; measured availability will determine whether the product is useful.