Batteries' Stress Test Triumph

When PJM's grid hit 162,000 MW on June 24—its third-highest peak ever—the performance gap between battery storage and thermal generation stopped being theoretical. Batteries discharged at 94% of rated capacity during the critical 4-to-8 PM window. Combined-cycle gas plants managed 87%. Simple-cycle turbines limped to 81%.

The implications ripple far beyond one heat wave. Capacity market constructs across PJM, ISO-NE, and MISO rest on the assumption that nameplate ratings minus historical outage rates predict real-world availability. Summer 2025 exposed that assumption as dangerously optimistic—precisely when reliability matters most.

The Physics of Performance Under Pressure

The thermal-storage performance gap isn't random. It reflects fundamental engineering differences that become acute during grid stress.

Gas turbines suffer efficiency losses as ambient temperatures rise. A combined-cycle plant rated at 500 MW under ISO conditions might deliver 460 MW when outside air hits 105°F. Batteries face no such derating—lithium-ion cells maintain output regardless of external temperature, provided thermal management systems function.

ERCOT's August price spikes illustrated a second divergence: response speed. When real-time prices hit $5,000/MWh—consistent with the grid operator's Operating Reserve Demand Curve pricing during scarcity—batteries responded within 4 seconds of dispatch signal. Gas peakers required 10 to 15 minutes to reach full output. That lag cost the system an estimated 1,200 MW during the critical 6-to-7 PM hour, when every megawatt commands premium value.

CAISO's longer operational dataset reinforces the pattern. The grid operator's 2024 Annual Report documented battery forced outage rates of 2.1% against 6.8% for gas-fired peakers. Four-hour duration batteries maintained full output for 97% of dispatched hours—performance that would have seemed aspirational five years ago.

Batteries Outperformed Gas During Peak Stress

Batteries Outperformed Gas During Peak Stress. Source: PJM June 24 peak event data. Batteries delivered 94% of rated capacity during the critical 4-8 PM window, outperforming thermal generation by 7-13 percentage points.

The Accreditation Reckoning

Current capacity market rules in PJM and ISO-NE credit thermal generators at nameplate capacity minus historical forced outage rates. The methodology implicitly assumes that past performance during normal operations predicts future availability during emergencies.

Summer 2025 demonstrated otherwise. Forced outage rates during extreme conditions exceeded historical averages by 2 to 4 percentage points across multiple thermal fleets. Units that cleared capacity auctions based on 95% availability delivered 85% when the grid needed them most.

California pioneered an alternative. CAISO's Effective Load Carrying Capability methodology calculates each resource's marginal contribution to system reliability based on probabilistic modeling. Resources that perform well during loss-of-load-probability hours receive proportionally higher credit.

The results diverge sharply from nameplate-based approaches. Under ELCC, four-hour batteries in CAISO receive 85% to 95% capacity credit. Gas peakers receive 75% to 85%. The differential reflects demonstrated performance differences—not theoretical ratings.

Who Wins, Who Loses

For battery developers, the performance data supports aggressive capacity bidding strategies. The operational record now demonstrates the reliability that earlier projects could only promise. Capacity revenue certainty improves with higher forced availability, and batteries face lower penalty exposure under capacity performance regimes.

Response Speed: Seconds vs. Minutes

Response Speed: Seconds vs. Minutes. Source: ERCOT August dispatch data. Battery response in 4 seconds captured full $5,000/MWh spike pricing; gas peakers' 10-15 minute ramp-up missed the peak entirely.

Energy margin capture during price spikes rewards fast response. Batteries capture the full $5,000/MWh spike price; slow-starting gas plants miss the peak entirely. With ERCOT's energy-only market structure offering no capacity payments to fall back on, real-time performance becomes existential.

Gas peaker owners face the inverse calculation. Assets that cleared capacity auctions at full accreditation now confront potential derating of 10 to 20 percentage points under ELCC-based frameworks. For a 200 MW combustion turbine, that's the difference between $8 million and $6.4 million in annual capacity revenue—assuming the methodology changes clear regulatory review.

Merchant generators with mixed portfolios have already begun repositioning. Several announced plans to co-locate battery storage at existing gas plant sites, capturing transmission interconnection value while hedging accreditation risk.

The Regulatory Pipeline

FERC Docket EL24-99 addresses PJM's capacity performance construct directly. Intervenors argue that current non-performance penalties fail to incentivize reliability improvements. Summer 2025 data strengthens that argument considerably—penalties that seemed adequate based on historical outage rates may prove insufficient when stress-event performance diverges this sharply from baseline.

The docket also examines whether storage should receive enhanced capacity credit. PJM currently caps storage accreditation based on discharge duration without considering superior forced outage performance. Reform proposals would base accreditation on demonstrated ELCC contribution—a methodology increasingly viewed as the model for other ISOs.

97%

97%. Four-hour duration batteries maintained full output for 97% of dispatched hours in CAISO—performance that would have seemed aspirational five years ago.

ISO-NE's pending capacity market reform under Docket ER25-456 proposes similar changes. The grid operator's analysis shows storage resources providing 1.4 times the reliability value per installed MW compared to combustion turbines. If adopted, New England's capacity market would shift billions in annual payments toward storage and away from aging gas fleets.

Meanwhile, FERC itself operates with reduced bandwidth. Commissioner Christie's August 2025 departure leaves the agency with just three members—complicating the path for any controversial rulemaking on capacity accreditation methodology.

What Comes Next

Capacity accreditation reform proceedings will accelerate through 2026, with PJM, ISO-NE, and MISO each examining ELCC-based methodologies. The summer 2025 performance data transforms what was an abstract modeling debate into a concrete reliability argument.

Thermal generation owners will resist changes that reduce their capacity values. Expect litigation over methodology changes to extend into 2027 or beyond. But the direction of travel seems clear—capacity markets will increasingly reward demonstrated performance over nameplate ratings.

The deeper question is whether current capacity constructs can adapt fast enough. Battery storage additions across major ISOs exceeded 12 GW in 2024 alone, up from 5.4 GW the prior year. Each summer stress event generates performance data that strengthens the case for accreditation reform. Thermal incumbents face a narrowing window to shape the rules before the rules reshape their economics.