How Much Can Demand Response Programs Save Utilities in Avoided Peak Costs?

Peak demand has always been expensive, and progressively more so every year. As electrification, increasingly common extreme and volatile weather, and data center growth push loads higher, utilities are under pressure to find capacity fast without breaking rate cases wide open. Demand response gives utilities a lever to pull before signing off on a new peaker plant or a costly substation upgrade.

Demand response works.

Data from 2023 to 2024 shows that demand response participation provided a total of 33,272 MW shifted, which could  have met roughly 6.5% of wholesale market peak demand for all RTOs and ISOs that year alone. In fact, research indicates that deploying demand response programs to shift a mere 5% of load could save more than $3b in costs annually.

The question then isn’t whether or not it’s worthwhile, but how much it saves, and under what conditions. Read on to learn what national studies and utility filings show about avoided peak costs, how those savings are calculated, and where the biggest opportunities sit today.

 

In This Article

• Why peak demand is so costly to serve
• What the data shows about demand response savings
• How utilities calculate avoided peak costs
• Why the math is shifting in demand response’s favor
• Frequently asked questions
• Glossary of terms

 

Why Peak Demand is so Costly to Serve

A utility’s system doesn’t need to be sized for the average hour of the year — it needs to be sized for the worst hour. That mismatch is where most avoidable costs live.

• Roughly 10% of the U.S. electric system exists to meet demand in about 1% of the year’s hours, according to research cited by ACEEE.
• Capacity built to cover those rare peak hours sits idle the rest of the year, but ratepayers still cover the fixed cost of building and maintaining it.
• New gas peaking capacity, historically the default answer to peak growth, has become significantly more expensive. According to reporting from the American Public Power Association (APPA), combined-cycle gas turbine projects cost $2,000 per KW, outpacing previous projections.
• Capacity auctions have reflected the same tightening. PJM’s most recent auction saw capacity prices hit the price gap, as retiring generation, construction delays, and load growth collided.

Every megawatt of peak that doesn’t need to be served by new infrastructure is a megawatt of avoided capital cost, avoided financing cost, and avoided rate pressure.

 

What Data Shows About Demand Response Savings

Utilities have run demand response programs for decades, long enough to generate real performance data, not just projections.

• Research indicates that demand response and energy efficiency programs can reduce U.S. electrical consumption by 8% by 2040. Furthermore, analysis suggests that the U.S. grid has nearly 60-200 GW of load flexibility potential over the next decade, providing opportunities to enhance load flexibility initiatives.
• Studies have found that for every dollar spent reducing peak demand, $2 to $3 are returned in consumer savings.
• The Federal Energy Regulatory Commission’s 2024 Assessment of Demand Response and Advanced Metering tracks potential peak demand savings by census division using EIA data, showing that potential savings shift year to year based on program design, enrollment, and how aggressively utilities call events.
• The gap between potential and actual savings is relatively consistent across these reports; programs that go uncalled or under-enrolled leave savings on the table. A utility long on capacity has less incentive to dispatch demand response fully, which is part of why realized savings often trail modeled potential.

The takeaway for planners: demand response’s value isn’t fixed. It’s a function of scale and strategy, including enrollment, dispatch frequency, and how well a program is designed to match the hours that actually drive system cost.

 

How Utilities Calculate Avoided Peak Costs

Avoided cost isn’t a single number — it’s a stack of separate benefit categories, and where the value concentrates depends heavily on the utility’s system conditions. This includes:

• Avoided generation capacity — the cost of the power plant (typically a gas peaker) the utility doesn’t have to build or contract for. This is usually the largest single component. It’s worth noting that virtual power plants (VPPs), a demand flexibility strategy similar to demand response, cost 40-60% of conventional generation alternatives.
• Avoided energy costs — lower wholesale energy purchases during high-price peak hours.
• Avoided transmission and distribution (T&D) investment — deferring or downsizing substation and feeder upgrades in constrained areas.
• Ancillary services — value from providing grid balancing and reliability services.

The cheapest megawatt utilities can access is one that they don’t have to build. For example, a 2019 Brattle Group study estimated close to 200 GW of cost-effective load flexibility potential in the U.S. by 2030, which is worth more than $15 billion annually in avoided system costs, driven primarily by avoided generation capacity, followed by avoided energy costs, T&D deferral, and ancillary services.

The value is also highly location-dependent:

• Lawrence Berkeley National Laboratory’s (LBNLs) demand response potential research found that while the average value to the T&D system is around $25 per kW-year, the top 5% of areas with localized grid constraints can see values of $160–$300 per kW-year.
• That spread matters for program design — a flat, one-size-fits-all incentive structure will systematically undervalue demand response in the constrained pockets of a system where it matters most.

 

Why the Math is Shifting in Demand Response’s Favor

Two forces are pushing avoided-cost calculations higher across the industry right now:

• Load growth is accelerating. Data centers, electrification, and manufacturing reshoring are driving faster peak growth than utilities have planned for in decades.
• Supply-side alternatives are getting more expensive and slower to build. Aggressively rising turbine costs, supply chain lead times, and permitting timelines make demand-side resources comparatively faster and cheaper to deploy than new generation.

Utilities evaluating whether an incremental capacity need should be met with steel-in-the-ground or with a load flexibility program are increasingly finding that the latter pencils out faster, that Grid-Edge distributed energy resource management systems (DERMS) minimize grid congestion while providing an affordable alternative to the costs of modernizing aging grid assets.

 

FAQs

Does demand response actually reduce customer bills, or just utility costs? Both, over time. Avoided capacity and energy costs typically flow through to rates, and several studies have found net positive returns for ratepayers overall.

How is “avoided cost” different from a demand response incentive payment? The incentive is what a utility pays participating customers. Avoided cost is the broader system-level savings (generation, T&D, energy, ancillary services) that the utility captures by not having to build or buy an alternative resource. Incentive payments are typically a fraction of total avoided cost.

Why do actual savings often fall short of potential savings? Enrollment shortfalls, underused dispatch, and program designs that don’t target the highest-value hours or locations are the most common gaps identified in FERC and ACEEE data.

Is demand response still relevant now that battery storage costs have dropped? Yes — most utility and Brattle-style studies treat demand response and storage as complementary, not competing, resources. Demand response tends to be faster to scale and lower-cost per kW, while storage offers longer, more predictable duration.

Does the value of demand response vary a lot by location? Significantly. As the LBNL research above shows, value concentrates in areas with local grid constraints, which is why locational program design tends to outperform flat, systemwide approaches.

 

Glossary of Terms

• Demand response — programs and incentives that encourage customers to reduce or shift electricity use during periods of high system demand.
• Avoided cost — the cost a utility does not incur because a demand-side or supply-side resource offset the need for it (e.g., a new power plant or grid upgrade).
• Peaker plant — a power plant, usually natural gas-fired, built to run only during the highest-demand hours of the year.
• Load flexibility — the broader category of demand-side resources, including demand response, that can shift or modulate electricity use in response to grid needs.
• Ancillary services — grid support services (like frequency regulation and reserves) that help maintain reliable, balanced electricity supply.
• Transmission and distribution (T&D) deferral — postponing or downsizing infrastructure upgrades by reducing localized peak load.
• Distributed energy resources (DERs) — small-scale, often customer-sited energy assets (thermostats, water heaters, EV chargers, batteries) that can be coordinated to provide grid value.
• Virtual power plant (VPP) — an aggregation of DERs coordinated to act like a single dispatchable resource for the grid.

 

The Bottom Line

Avoided peak cost isn’t theoretical: it shows up in real utility filings, capacity auction prices, and national assessments year after year. The size of the savings depends on program design, enrollment, dispatch discipline, and how well a utility targets the specific hours and locations driving its costs. As load growth accelerates and new generation gets more expensive to build, that math is only becoming more favorable for demand-side approaches.