Demand response programs shifted 33,272 megawatts across U.S. wholesale markets in 2023-2024, covering roughly 6.5 percent of all RTO and ISO peak demand and creating an estimated $3 billion in annual avoided costs – a scale that now rivals new gas peaker construction without the capital lock-in. With capacity prices spiking in PJM’s latest auction and combined-cycle gas turbine costs hitting $2,000 per kilowatt, utilities facing load growth from data centers and electrification have a working alternative to rate-basing new peakers. The economics have flipped: every megawatt of peak reduced through flexible load is a megawatt of generation, transmission, and distribution capacity that never needs to be built, financed, or recovered from ratepayers.
The Structural Mismatch Driving Peak Costs
The U.S. grid is built for the worst hour, not the average hour. Research cited by the American Council for an Energy-Efficient Economy (ACEEE) shows that approximately 10 percent of the entire electric system exists solely to serve demand during about 1 percent of annual hours. That overbuild sits idle 99 percent of the time, yet ratepayers cover its fixed costs – capital recovery, financing, operations, maintenance, and property taxes – year-round. Historically, utilities met this need by adding gas-fired peaking units, which run infrequently but provide firm capacity when called upon.
That default solution has become dramatically more expensive. The American Public Power Association (APPA) reports combined-cycle gas turbine projects now cost $2,000 per kilowatt, exceeding earlier projections. At the same time, capacity markets are signaling scarcity: PJM’s most recent capacity auction cleared at the price cap, reflecting a collision of retiring thermal generation, interconnection queue delays for new resources, and accelerating load growth. When capacity prices hit the administrative ceiling, the market is effectively saying it cannot procure enough firm resources at any reasonable cost. That signal alone reframes demand response from a “nice-to-have” efficiency measure into a critical capacity resource.
The avoided-cost calculation is straightforward but often undercounted. A megawatt of peak reduction avoids not just the combustion turbine itself – roughly $1.5-2 million per MW installed – but also the associated transmission upgrades, distribution reinforcement, interconnection studies, permitting timelines, and the carrying charges on all of that capital over a 20-30 year book life. When a utility defers or cancels a peaker project, it also avoids the rate-case fight to recover those costs. That political and regulatory friction has real value, even if it never appears in a standard levelized cost of energy (LCOE) comparison.
Why the Math Is Shifting Toward Flexible Load
Three converging trends are making demand response economics more compelling than at any point since the first utility load-control programs in the 1980s. First, the resource base is expanding beyond traditional industrial curtailment. Smart thermostats, water heater controls, EV charging management, and behind-the-meter battery aggregation now allow utilities to tap residential and commercial load at scale. The 33,272 MW figure from 2023-2024 already reflects this diversification; it is not solely large-factory interruptible tariffs.
Second, the marginal cost of acquiring incremental demand response is falling. A utility launching a bring-your-own-thermostat (BYOT) program today pays primarily for enrollment incentives, platform integration, and measurement and verification – costs that scale sub-linearly with participation. By contrast, each additional megawatt of gas peaker capacity requires a near-linear increase in capital expenditure, land, gas pipeline access, and permitting risk. The crossover point where DR becomes cheaper per firm MW than new thermal capacity has already been reached in many jurisdictions, especially when the social cost of carbon or local air quality compliance is factored in.
Third, the value stack for demand response is deepening. In organized markets, DR resources can now stack capacity payments, energy arbitrage, ancillary services (regulation, spinning reserves), and transmission congestion relief. A residential battery fleet enrolled in a virtual power plant (VPP) program in California or New York may earn revenue from Resource Adequacy, day-ahead energy markets, and frequency regulation simultaneously. That multi-stream revenue improves the participant’s payback and the program administrator’s cost-effectiveness ratio. The $3 billion annual savings estimate cited in national studies assumes only a 5 percent load shift – a conservative target given that some VPP pilots have demonstrated 10-15 percent peak reduction on targeted feeders.
That points to a structural shift: demand response is no longer a demand-side management (DSM) line item buried in an energy efficiency docket. It is increasingly procured through the same capacity auction mechanisms as generation, evaluated on equivalent $/MW-day terms, and dispatched by grid operators with the same visibility as a peaker plant. The distinction between “supply-side” and “demand-side” capacity is eroding in market rules, and utilities that treat DR as a procurement category – not a conservation program – are capturing the full avoided-cost value.
Who This Affects
- Utility resource planners: Re-run integrated resource plans (IRPs) with demand response modeled as a selectable capacity resource at current enrollment costs, not as a fixed DSM budget line. The 33 GW national figure suggests 50-100 MW of cost-effective DR potential exists in most mid-sized service territories.
- Storage and VPP developers: Stack capacity market revenue with distribution-level services (feeder peak shaving, voltage support) to improve project IRRs; utilities are issuing RFPs for non-wires alternatives that explicitly value locational DR.
- State public utility commission staff: Require utilities to quantify avoided transmission and distribution (T&D) costs in DR cost-effectiveness tests, not just avoided generation capacity. The $2,000/kW gas peaker benchmark understates total system savings when T&D deferral is included.
- Large commercial and industrial customers: Evaluate enrollment in utility DR programs or third-party aggregator contracts as a hedge against rising capacity charges – especially in PJM, NYISO, and ISO-NE where capacity tags are set by coincident peak hours.
What to Watch Next
- PJM and NYISO capacity auction results through 2026: Sustained high clearing prices will accelerate utility DR procurement targets and may trigger new state mandates for demand-side resource acquisition.
- FERC Order 2222 implementation compliance filings: Watch for RTO/ISO rules that allow aggregated DR to participate directly in wholesale markets without utility intermediation – this could unlock an additional 10-20 GW of behind-the-meter flexibility by 2030.
- Data center load interconnection agreements: Hyperscalers negotiating large-load tariffs increasingly offer flexible load as a condition of service; track how many MW of data center demand become dispatchable DR resources.
- Advanced metering infrastructure (AMI) 2.0 deployments: Next-generation meters with edge computing enable sub-second load control and measurement, reducing DR verification costs and expanding eligible end-uses to include EV fleets and heat pumps.
Bottom line: The 33 GW of demand response already operating in wholesale markets is not a pilot – it is the largest virtual peaker fleet in the country, built without a single rate case for new generation. Utilities that procure the next 5 percent of peak load as flexible demand instead of gas turbines will avoid billions in capital, reduce carbon exposure, and keep rates flatter for the 99 percent of hours when the system isn’t peaking.
Read the full report at Energy Central
Note: facts and figures attributed above to reflect that outlet's original reporting. Broader context, cross-sector connections, and forward-looking scenarios reflect independent analysis by our editorial team.
About this article: Drafted by Energy Ai with AI-assisted research and writing based on public reporting, then reviewed under our editorial process before publication.
Leave a Reply