What is time to power, and why does it decide which data-centre sites get built?
Time to power is how long it takes for firm electricity to reach a site’s load. In 2026 it is the binding constraint on US data-centre development: sites can be built in months, but grid interconnection queues now run several years. A site with a 24-month behind-the-meter pathway is worth far more than the same site waiting four years for the grid.
What is the binding constraint on US electricity in 2026?
Time. According to the Earthshot Foundation’s June 2026 paper, synthesising a year of dialogue with more than 200 grid practitioners, the binding constraint is no longer generation or capital but the time required to interconnect new load. Data centres are built in months; interconnection queues now run several years and contain roughly three times the country’s total installed operating capacity.
Does going behind the meter solve it?
No — it changes the shape of the constraint. Off-grid, the bottleneck is no longer the interconnection queue. It becomes a matching problem: how fast a generation technology, a site and the capital willing to underwrite both can find and qualify each other. Today that search typically takes 18–24 months — the same order of magnitude as the grid queue it was meant to escape.
Which behind-the-meter technology has the fastest time to power?
Natural gas reciprocating engines, at roughly 12–24 months from order to commercial operation once the gas connection and air permit are secured.
Why does time to power change what a site is worth?
A 500 MW load with a four-year interconnection queue is worth a fraction of the same load with a 24-month behind-the-meter pathway. The gap between those two values is effectively a new asset class: power that can be delivered inside an offtaker’s deployment window.
More answers
Matching technology, sites and capital
SitePower.ai is an application-only marketplace that qualifies each side of an energy infrastructure deal and brings the right counterparties together.