The American electric-power sector is now confronting two opposing forces at once: the largest load-growth cycle in a generation and a federal policy shift that is pulling back the incentives, financing structures, and permitting pathways expected to deliver much of the next wave of supply. Demand is rising from artificial-intelligence data centers, cryptocurrency operations, manufacturing reshoring, and the electrification of transportation and buildings. At the same time, new federal constraints are weakening the economics of wind, solar, and storage projects that utilities and grid operators had expected to rely on.

A late-August 2026 analysis by the Natural Resources Defense Council, modeled by Evolved Energy Research, estimates that 390 to 540 gigawatts of planned wind, solar, and battery storage may not reach commercial operation over the next decade. For an industry focused on keeping supply ahead of demand, that is not a political footnote. It is a resource-adequacy problem that will shape interconnection queues, capacity prices, and customer bills well into the 2030s.

A Reversal Meets a Surge

The scale of the demand story is no longer in dispute. The U.S. Energy Information Administration projects national electricity consumption rising from roughly 4,195 billion kilowatt-hours in 2025 to about 4,269 billion in 2026 and 4,399 billion in 2027, setting new records in each year. For the first time, commercial-sector sales are expected to surpass residential sales, a structural shift that reflects how thoroughly hyperscale computing has rewired the load curve. U.S. electricity demand growth of this kind is especially difficult to serve because it is concentrated in large, always-on, interconnection-hungry facilities. It becomes more difficult still when the resources best positioned to meet that load on a short timeline are simultaneously losing policy support.

That is the essence of the current moment. The One Big Beautiful Bill Act, enacted in 2025, rewrote the economics of new wind and solar by compressing the technology-neutral credits under Sections 45Y and 48E. Instead of the gradual, emissions-based phase-down envisioned under the Inflation Reduction Act, the law creates an abrupt cliff: projects must begin construction on or before July 4, 2026, or be placed in service by December 31, 2027, to remain eligible. Missing either milestone means losing the credit entirely. New tariffs on imported components, tighter Treasury guidance narrowing what qualifies as the start of construction, and the federal buyback of several offshore-wind leases compound that shift. Taken together, the Inflation Reduction Act rollback removes much of the financial scaffolding that supported the project pipeline utilities and grid operators had been counting on.

The practical question is not whether new demand will materialize; it is which resources can be financed, permitted, interconnected, and delivered quickly enough to meet it. The answer will vary by region, but the planning challenge is national: policy risk is now becoming a resource-adequacy variable.

Quantifying the Capacity at Risk

The NRDC analysis is notable less for its advocacy framing than for its attempt to size the shortfall. Amanda Levin, the organization’s director of policy analysis, summarized the modeling bluntly: “We lose more than half of everything that we expected to be able to build.” The 390-to-540-gigawatt range spans wind, solar, and storage combined, measured against a decade-long baseline of expected additions. Even at the low end, subtracting nearly 400 gigawatts of planned capacity from a system already racing to serve record load represents a substantial tightening of reserve margins across nearly every region.

What makes the figure consequential for planners is the modeled replacement rate. According to the analysis, only about 9 gigawatts of additional natural-gas capacity would come online to substitute for the lost renewables. That is only a small fraction of the shortfall, and the reasons are familiar to anyone tracking the supply chain. Gas-turbine order books at major original-equipment manufacturers are stretched into the early 2030s, fuel-price volatility complicates long-term economics, and in many resource plans wind, solar, and storage remain the lowest-cost incremental megawatt-hours even without subsidy. The renewable energy capacity growth the market expected was not merely a climate ambition; it was the least-cost, fastest-to-interconnect answer to load growth. Removing it does not automatically summon an equivalent quantity of dispatchable generation. More often, it leaves a hole.

Why Gas Cannot Simply Fill the Gap

The comfortable assumption in some policy circles is that a pivot away from renewables is really a pivot toward the firm, dispatchable thermal capacity the grid needs anyway. The reality on the ground is more constrained. Heavy-duty combined-cycle turbine deliveries are already booked years into the future, so a developer deciding today to build gas instead of solar cannot expect comparable speed. In some cases, that pressure is nudging projects toward less-efficient simple-cycle units that can be sited faster but burn more fuel per megawatt-hour and emit more. That outcome does not solve the timing problem cleanly, and it raises the emissions intensity of the marginal megawatt.

There is also a timing mismatch that matters for reliability. Data-center loads are arriving in large blocks over eighteen-to-thirty-six-month horizons. Solar and storage, permitting and interconnection aside, can be manufactured and energized on a schedule that roughly matches that cadence. New gas, by contrast, must contend with turbine queues, pipeline capacity, and multi-year construction. When the fastest resource class is removed from the supply stack, the system does not simply substitute the next resource; it waits. For grid operators managing tightening reserve margins, that waiting period is where grid reliability and resource-adequacy risk concentrates.

The Affordability and Reliability Math

Cost is the dimension most likely to reach customers and regulators first. The NRDC modeling projects average household electricity rates rising an additional 4.2 to 5.5 percent by 2035 relative to a trajectory that preserved the prior incentive structure. It also projects roughly $5 billion to $15 billion in additional annual fossil-fuel spending and about $45 billion less in tax incentives flowing to the sector. The organization frames the cumulative effect as roughly $700 billion in forgone clean-energy investment by 2035, with power-sector carbon emissions potentially doubling over the same period. Reasonable analysts will debate the precise magnitudes, which depend heavily on gas-price paths, load realization, and how aggressively states backfill federal policy. But the direction is consistent with basic supply economics: constrain the lowest-cost incremental resource during a demand surge, and the clearing price of electricity tends to rise.

For a professional audience, the more actionable insight is regional. The states with the largest data-center pipelines and the thinnest reserve margins are where a supply pullback will bite hardest, and where capacity-market and bilateral-contract prices are likely to register the strain first. Utilities filing integrated resource plans over the next two cycles will have to reconcile load forecasts that keep climbing with a supply menu that has narrowed. Some, including Duke Energy in the Carolinas, continue to plan substantial solar additions well into the 2040s, a reminder that the underlying economics still favor renewables where they can be permitted and financed. The open question is how many planned megawatts survive the credit cliff, the tariff overhang, and the tighter construction-start rules.

The Counterview: Dominance, Durability, and Market Pull

An honest assessment has to give the opposing case its due. The administration’s energy-dominance agenda rests on a coherent, if contested, premise: that a reliable, affordable grid is best anchored by firm, domestically fueled generation, and that years of subsidy-driven renewable growth left too much of the system dependent on weather-variable output backed by ratepayer-funded incentives. Proponents point to record overall generation, an expanding role for nuclear that includes a policy push on new reactors, and a reaffirmed place for the existing coal fleet as evidence that the lights will stay on. They argue that removing subsidies simply exposes renewables to a fair market test, and that genuinely economic wind and solar will be built regardless. There is empirical support for part of this view: reporting through 2026 has documented renewables continuing to grow despite the policy headwinds, because their levelized-cost and speed advantages have not disappeared.

Skeptics of the NRDC framing also note that advocacy modeling tends to hold demand and technology costs on trajectories that maximize the estimated loss, and that a 390-to-540-gigawatt range is wide enough to signal real uncertainty. Federal and state actors continue to move on the supply problem through other levers. In August 2026, FERC approved a Southwest Power Pool topology-optimization plan aimed at squeezing more usable capacity out of existing wires, an example of the kind of operational efficiency that can partly offset slower resource additions. The debate, in other words, is not whether the grid will function, but at what cost, at what emissions intensity, and with how much margin for error during the extreme-weather events that increasingly define reliability. Those who defend the policy shift and those who warn against it are ultimately arguing about the size of the buffer between supply and demand, not the existence of the challenge.

Implications for Planners and Investors

For the practitioners who will live with the consequences, several implications follow regardless of where one lands politically. First, the value of speed has risen. In a market where the fastest-to-build resource class faces a policy cliff, any megawatt that can interconnect quickly, whether behind-the-meter generation, storage paired with existing sites, or brownfield repowering, commands a premium. Second, contract certainty matters more than ever. Large-load customers and developers are moving toward long-term structures that lock in supply, and the counterparties who can offer firm delivery will shape the next wave of deals. Third, the transmission and operational-efficiency toolkit deserves fresh attention, because incremental capacity wrung from existing infrastructure is now competing directly with new generation that has become slower and costlier to add. Notably, the same period saw the Department of Energy cancel several designated national transmission corridors, underscoring that the wires side of the equation is contested as well.

The strategic reality is that the industry is being asked to serve record demand with a narrower supply menu and fewer subsidized options. That does not make the task impossible, but it does raise the stakes on execution. The clean energy tax credit repeal has not ended the buildout; it has changed which projects pencil out, on what timeline, and at whose expense. Planners who treat the NRDC headline number as a precise forecast will miss the point. Treated instead as a stress test, the 500-gigawatt figure is a warning about how quickly planning margins can evaporate when policy and demand move in opposite directions.

Conclusion

The defining tension of the U.S. power sector in late 2026 is no longer whether demand will grow; every credible forecast agrees that it will, at a pace unseen in decades. The tension is whether the supply side can respond when the resources best suited to rapid deployment are losing their policy tailwinds at the exact moment they are most needed. The NRDC analysis, whatever one makes of its assumptions, crystallizes that tension into a number large enough to command attention across boardrooms, control rooms, and regulatory dockets. Firm generation advocates and renewable developers will continue to contest the framing, and the true outcome will depend on gas-price paths, state-level policy responses, and how much capacity survives the 2026 and 2027 deadlines. What is not in doubt is that the margin for error is thinning. For an industry whose core mandate is to keep supply ahead of demand at just and reasonable rates, the widening gap between record load and constrained resources will define planning, pricing, and reliability for years to come.