What does speed to power mean?

Speed to power is the elapsed time between committing to a data center site and having utility power energized and available for IT load. It has become the primary basis of competition in data center development because power availability, not construction, is the binding constraint.

Why is speed to power the deciding factor in site selection?

Because compute demand is immediate while grid capacity is queued. A site that energizes 18-24 months sooner than an alternative lets a tenant capture revenue and deploy hardware ahead of competitors, which is worth more than almost any other site attribute.

How does the distribution path improve speed to power in Texas?

Sites whose aggregate demand stays under 75MW are generally served through the local utility's distribution process — outside ERCOT's large-load framework and the additional requirements Texas SB6 applies to 75MW+ sites — subject to the utility's service rules and local system capacity. The utility process runs on capacity checks, service applications, and engineering studies rather than multi-year regional transmission studies.

Can sub-75MW sites scale beyond 75MW later?

Within limits. Sites can be engineered for phased expansion up to the 75MW aggregate site threshold — the large-load regime is triggered by total demand at the site, not by the size of any single phase. Growth beyond that total means entering the large-load framework deliberately, or adding capacity as separate sites. Expansion strategy should be engineered into the site design from day one.

What role do BESS and behind-the-meter generation play?

Behind-the-meter generation can carry early load while utility service ramps; battery storage adds peak management, ride-through, and power quality — both complements to, not substitutes for, a utility interconnection position.

Insights · 02 — Timing

Speed to power.

Time-to-energization has replaced price per acre as the number that decides data center transactions. Here's why, and how the timeline actually breaks down.

A tenant deploying compute measures cost in idle capital. Racks, cooling and networking are ordered before ground breaks; every month the site isn't energized, that equipment depreciates without producing revenue. At current AI-infrastructure economics, a year of delay can outweigh the entire land basis of the site.

This is why powered land transacts on go-live date, not acreage — and why the interconnection path a site takes is the single most consequential engineering decision made on its behalf.

Where the years go

A transmission-connected large load enters a queue: screening study, full interconnection study, facilities study, then construction of whatever network upgrades those studies require. Each stage has its own clock, each is sequential, and the upgrade construction at the end — new lines, breakers, transformer additions — is measured in years and often shared with other queue positions that can restart the math when they drop out.

A distribution-connected site under the large-load threshold skips that queue entirely. The utility studies the load against existing distribution capacity, extends facilities where needed, and energizes — a process measured in months. TPDS compresses it further by doing the work before the tenant arrives: our sites carry executed IAs and commissioned substations, so the 8–12 week figure covers tenant-specific configuration and final utility steps, not discovery.

What the range depends on

The honest caveats: metering configuration, tenant load profile, and utility crew scheduling all move the needle inside the range. Ranges reflect published ERCOT large-load interconnection timelines and TPDS engineering targets for distribution-connected sites; site-specific timing is confirmed in diligence. For the full step-by-step comparison of the two paths, see ERCOT Interconnection →