Power Boom Forecast Hides a Construction-Readiness Gap
A new forecast from MSI Economics, the research arm of construction technology firm MOCA Systems Inc., puts a number on something field contractors have suspected for a while: the AI-driven power buildout headlines are running well ahead of what’s actually going to break ground, Engineering News-Record reports.
The firm’s report, “Recharging America: The Construction Response to Rising Power Demand,” projects U.S. power-generation construction spending could roughly double to $150 billion annually by 2030, with nearly $691 billion in cumulative spending through the decade. But the report’s real value for contractors isn’t the topline number. It’s the methodology MSI used to separate announced capacity from construction-ready capacity, and the gap it found is significant.
Background
MSI built its model using NERC regional reliability data, EIA capacity benchmarks, GlobalData project-pipeline tracking, and interconnection research from Lawrence Berkeley National Laboratory. To count a 2026 or 2027 project as executable rather than speculative, MSI required four things: a cleared or advanced interconnection-queue position, an executed interconnection agreement, reserved transformers, turbines and switchgear, and committed financing. Miss any one of those, particularly equipment reservations, and the report pushes the project’s start into 2028 or later.
The results are stark. Of 12 GW to 16 GW of data-center-related capacity projected for 2026 delivery, MSI found only about 5 GW is currently under construction. Large power transformer lead times have stretched from a pre-2020 norm of 24 to 30 months to as long as 128 weeks, generator step-up transformers can run 144 weeks, and heavy-frame gas turbines for combined-cycle plants are being quoted at five to seven years from order to commercial operation. Interconnection timelines add another layer: MSI found projects that reached service from 2022 through 2025 spent a median 46 months in Texas’s queue and nearly 100 months in California’s.
AGC chief economist Ken Simonson told ENR he expects strong demand growth for power and grid upgrades but couldn’t independently verify MSI’s spending forecast. He and ABC chief economist Anirban Basu both flagged labor as a separate constraint, with Basu noting electrician wages rose 7.2% over the past year amid what he called “particularly outsized” demand from data centers and power projects relative to supply.
Analysis
The headline $150 billion figure is going to circulate widely, and it’s going to get used to justify aggressive bidding, crew mobilization and equipment pre-buys across the sector. That’s exactly where this report should give subcontractors pause rather than confidence.
MSI’s own numbers show the announced pipeline and the buildable pipeline are two different things, and the gap is measured in years, not months. A developer with 500 MW of announced capacity and a signed data center lease is not the same as a developer with a cleared interconnection agreement, reserved switchgear and locked financing. The report’s four-condition test exists precisely because too much of the current “boom” narrative treats those as interchangeable.
The equipment bottleneck is the most concrete signal here. Turbine and transformer lead times measured in years mean that even fully financed projects announced today may not see steel in the ground until 2028 or beyond, regardless of how much capital is behind them. And MSI’s own relief scenario, where equipment supply loosens starting in 2028, is conditional on manufacturers actually following through on announced capacity expansions and enough projects surviving long enough to keep those orders in place. Grain-oriented electrical steel, a core transformer input with only one U.S. producer and roughly 80% import dependence, is a single point of failure sitting underneath that entire recovery timeline.
Geographically, the money is concentrated. ERCOT, PJM and MISO together account for about 87% of MSI’s implied construction demand through 2030, with Texas alone tied to $76.2 billion in generation construction as peak demand there is projected to grow 46.8%. Solar makes up roughly 74% of projected capacity additions, natural gas is the only dispatchable source scaling meaningfully at about $29.4 billion annually by 2030, and nuclear contributes little near-term volume unless MSI’s more speculative small modular reactor scenario materializes.
The behind-the-meter trend adds another wrinkle. Howden’s separate data center risk report projects on-site generation could supply about 25% of U.S. hyperscale data center power by 2030, roughly 10 GW. Oracle’s Project Jupiter campus in southern New Mexico, which swapped a planned gas-turbine and diesel setup for a Bloom Energy fuel-cell microgrid, shows developers are already routing around grid and equipment delays rather than waiting them out. That shifts construction scope and specialty trade demand toward on-site generation packages rather than traditional utility-scale interconnection work.
What It Means for Subcontractors
- Before committing crews or reserving equipment against a “data center power project,” confirm the four MSI benchmarks apply: cleared interconnection queue position, executed interconnection agreement, reserved transformers/turbines/switchgear, and locked financing. Projects missing any one are more likely pushed to 2028 or later.
- Electrical and E&I contractors bidding transformer or switchgear-dependent scopes should price in lead times of up to 128 weeks for large power transformers and 144 weeks for generator step-up units, and structure contracts to protect against schedule slippage while equipment sits in backorder.
- Gas turbine-dependent mechanical and pipefitting trades should note five-to-seven-year order-to-commercial-operation timelines for heavy-frame turbines when sequencing crew commitments on combined-cycle plant work.
- Prioritize bid pursuit in ERCOT, PJM and MISO territories, which account for about 87% of MSI’s projected construction demand through 2030; Texas alone is tied to $76.2 billion in generation construction through the decade.
- Solar EPC and civil crews should expect the largest volume of near-term work, since solar represents roughly 74% of projected capacity additions, while gas-focused trades should track the $29.4 billion annual spending level MSI projects by 2030.
- Electricians and skilled trades should factor in wage pressure now, with average hourly electrician earnings up 7.2% over the past year amid competition from data centers, LNG facilities and semiconductor fab work, per ABC’s Anirban Basu.
- Contractors eyeing behind-the-meter or microgrid scopes, like the fuel-cell and renewable generation packages tied to Oracle’s Project Jupiter in New Mexico, should track Oracle’s Sept. 8 solicitation for 2 GW of solar, wind and geothermal capacity targeted for delivery between 2027 and 2031.
- Firms in California should account for interconnection timelines running close to 100 months based on recent completions, versus a 46-month median in Texas, when forecasting when queued projects will actually convert to construction spend.
