The grid was designed for yesterday’s load.

Data centers are changing the scale, concentration, and timing of electricity demand. The investment question is no longer simply how much generation to add—it is where the next constraint will move.

Read the findings ↓

1.9%

Forecast U.S. electricity-load growth in 2026.

EIA ↗

2.5%

Forecast load growth in 2027.

EIA ↗

4.4%

Estimated data-center share of U.S. electricity use in 2023.

DOE / LBNL ↗

6.7–12%

DOE/LBNL projection range for data-center share in 2028.

DOE / LBNL ↗

Executive finding

Demand is accelerating faster than infrastructure cycles.

EIA expects U.S. load growth to continue through 2027, with particularly strong growth in regions that include large data-center markets. Engineer View’s interpretation is that aggregate energy supply is only the first-order issue. The commercially decisive constraint may be a transformer, substation, interconnection study, transmission path, permitting sequence, or construction workforce.

Key takeaways

Concentrated loads make location and connection timing as important as annual energy.
Constraints migrate; solving generation does not solve delivery or interconnection.
Robust decisions preserve optionality across several credible demand paths.

01 · The shape of load

A megawatt is not a sufficient description.

Planners need the load’s arrival date, ramp rate, duty cycle, flexibility, redundancy requirement, and geographic concentration. Large computing campuses can arrive in increments that are material to a local balancing area even when the national percentage appears modest.

Figure 1. U.S. electricity demand has shifted from a low-growth regime.

2005–20190.1% / yr

2020–20251.7% / yr

2026 forecast1.9%

2027 forecast2.5%

Source: U.S. Energy Information Administration, March 2026. Values are not a continuous series; historical compound growth is shown alongside annual forecasts.

02 · Constraint migration

The critical path moves through the system.

When demand accelerates, the binding limit can travel across a connected chain. Adding capacity at one stage exposes the next bottleneck. The practical unit of analysis is therefore the full connection pathway.

Generation→Interconnection→Transmission→Substation→Site

System adequacy is a coordination problem disguised as a capacity number.

03 · Decision scenarios

Plan for triggers, not false precision.

Managed growth

Load arrives in phases; flexible demand and existing headroom absorb early growth.

Prioritize staged capacity and commercial triggers.

Connection squeeze

Projects cluster faster than local delivery assets and studies can advance.

Prioritize site optionality and queue strategy.

System acceleration

Compute, manufacturing, and electrification reinforce one another.

Prioritize long-lead equipment and corridor capacity.

These are decision scenarios, not probability-weighted forecasts. They are designed to reveal which actions retain value when the demand path changes.

04 · Implications

Time-to-power becomes a strategic variable.

Developers should treat energization as a portfolio of dependent milestones, not a promised date. Utilities need clearer visibility into speculative versus committed load. Investors should test whether project economics survive connection delay, phased energization, and higher-cost interim supply.

Decision test

If energization moves by 18 months, which option preserves the most value?

Alternative site or service territory
Phased load and modular equipment
Demand flexibility or behind-the-meter capacity
Commercial milestones tied to utility evidence

Sources & method

Research record.

This report synthesizes current public statistics and Engineer View systems analysis. Quantitative claims are linked to primary government sources; scenario structure and implications are Engineer View interpretation.

General research only. Not engineering certification, project-specific advice, or investment advice. Data may be revised by the publishing agencies.