Two builds, one clock: how data center time-to-power may reshape E&C

Data center construction in the US can finish years before power arrives. Boards should measure that exposure now.

In brief

• For Engineering & Construction (E&C) firms, a completed private-side asset can sit idle for years, earning nothing until it is connected to power.
• Federal reform of large-load interconnection is under way. The power plant queue is the closest proxy for data centers, with a 2025 median wait over 5 years.
• Integrators can't control utility timing, so contracts should cap the delay risk they take on. Boards should measure power-linked backlog and stage investment.

Why data center time-to-power is now an E&C problem

In the EY current-trajectory model, US electricity demand rises about 12% from 2026 to 2030 and a further 14% from 2030 to 2035, with data centers accounting for about 49% of that later increase.[2] In one stylized model of a fully grid-connected 1 GW facility with roughly $38 billion in up-front capital, the owner’s own substation works account for less than 0.5% of the total. The date the facility is energized sets when anything it is built to hold can start earning.[3]

Project owners carry most of this schedule exposure today. Contractors absorb a growing share of it; in backlog work they may have to resequence or hold open until someone else’s power date arrives.

The grid development clock

Load interconnection is the process of connecting a large electricity consumer, such as a data center campus, to the grid. Generation interconnection is the parallel process for connecting new supply. Load interconnection has historically been a state and local matter, so there is no long-running national data set comparable to the generation queue. Federal rulemaking on the question began in October 2025, after the Secretary of Energy directed the Federal Energy Regulatory Commission (FERC) to consider an advance notice of proposed rulemaking on large-load interconnection.[4] On a separate track in June 2026, FERC issued show cause orders directing the six regional grid operators it regulates to justify or reform their large-load rules across five reform categories, one of which is transparency into transmission upgrade costs.[5] Until those proceedings conclude, generation queues remain the closest published proxy. In regions where duration data are available, Lawrence Berkeley National Laboratory reports that the median project reaching commercial operation in 2025 took more than five years from interconnection request to operation.[6]

Who owns the seam between the two builds?

Neither side of that seam can manage it alone, which is what makes the resulting delay expensive. That work is rarely contracted as an accountable, risk-bearing scope. Call the party that takes on this work a two-build integrator: a firm or consortium that sits between the campus build and the power build and takes contractual ownership of that interface. It can own everything on the private side, up to and including private generation. It cannot own the utility's timeline. That is the line between the delays an integrator can price and the ones that must remain relief events.

FERC is pressing a related question about who pays. In the same June 2026 orders, it asked each region to justify or add a pro forma cost recovery agreement for large loads. Such an agreement would assign the cost of a network upgrade to the customer taking service on behalf of the load that triggers the upgrade.[5] As that cost exposure becomes more defined, owners have more reason to push schedule risk downstream.

We see three models forming. Some contractors are adding power-side capability. Energy developers are leading where generation is the scarce asset, while large buyers are seeking control through co-located generation and long-term offtake. None of the three is settled practice. Few public examples show any of them working repeatedly at scale, and we found no like-for-like premium for integrated delivery in the public pricing reviewed for this analysis.

Boards must decide how far to move on the integrator role, and when. Our full point of view sets out where each delivery model is strongest, the boundaries of contractual control and how to tell a real risk-transfer premium from scarcity pricing.

Three questions for the board agenda

In a delivery-constrained market, a growing backlog is a risk indicator. Boards should be able to answer:

• How much of our backlog depends on a power date we do not control?
• Who owns the interface on our largest programs?
• Where in our contracts is energization accountability priced?

When private-side construction finishes before firm power arrives, capital is tied up and carrying costs accrue. Buyers may increasingly reward firms that can measure that exposure, price it, staff the integrator role and carry a bounded share of the schedule risk. E&C boards can start measuring before the rest of that is settled.

Mohit Ahuja and Gautam S Rashingkar contributed to this report and article.