PJM's frequency regulation market is the smallest of the major ancillary services markets in megawatt terms, but for fast-responding battery storage assets it remains one of the most lucrative revenue lines per MW installed. It's also one of the most misunderstood lines in battery storage underwriting decks.
This piece walks through the basics of how PJM regulation pricing works, the key differences between RegA and RegD signals, and what the latest performance scoring updates mean for how we underwrite storage projects in 2026.
What PJM is buying when it buys regulation
Frequency regulation is the second-by-second balancing service that keeps grid frequency at 60 Hz. PJM dispatches regulation through Automatic Generation Control (AGC) signals sent to enrolled resources every two seconds. Resources are paid for:
- Capability — being available to respond, paid in dollars per regulation MW per hour
- Performance — actually responding accurately, paid in dollars per "mileage" MW (the absolute MW change the resource was asked to make)
Importantly, both components are scaled by a resource's performance score — a measure of how accurately and quickly the resource tracks the AGC signal. A score of 1.0 is perfect; battery storage assets routinely score above 0.95.
RegA vs. RegD: what's the difference?
PJM dispatches two regulation signals:
RegA (traditional regulation)
A slower, longer-duration signal designed for traditional generators — gas turbines, hydro units, and some thermal plants. RegA energy balances out over time periods of 15+ minutes.
RegD (fast regulation)
An energy-neutral, faster signal designed for fast-responding resources like batteries. RegD signals are designed to net to zero over short periods, which makes them well-suited for resources with limited energy capacity (i.e., batteries).
For battery storage, RegD is almost always more attractive because:
- The signal is energy-neutral, so a battery isn't slowly drained over the hour
- Mileage is significantly higher for RegD, which means more performance revenue per MW
- Storage resources can score very high on RegD performance because they respond near-instantly
How the math actually works
A simplified hourly revenue equation for a regulation-enrolled battery storage resource:
Hourly revenue ≈ (RMCCP × Capability MW × Performance Score) + (RMPCP × Mileage MW × Performance Score)
Where:
- RMCCP = the regulation market capability clearing price ($/MW-hr)
- RMPCP = the regulation market performance clearing price ($/mileage-MW)
- Capability MW = the MW the resource offered into the market
- Mileage MW = the absolute MW change PJM asked the resource to make
Mileage for RegD is much higher than for RegA. Recent years have shown RegD mileage ratios in the 2.5–3.5× range — meaning a 1 MW battery providing RegD might be asked to move 2.5 to 3.5 MW of net dispatch every hour.
What's changed in performance scoring
PJM has refined performance scoring over the past several market years to better reflect actual signal-tracking accuracy. The most consequential changes for storage:
- Stricter tolerance bands for RegD performance — small misses now matter more
- Updated benchmarking that compares resource response to optimal response, not historical fleet response
- More aggressive penalties for repeated under-performance, with potential mileage adjustments
For well-controlled battery storage assets, none of these changes are existential. A modern battery with a properly tuned response control loop should still clear performance scores well above 0.95. But the changes do tighten the gap between best-in-class and average operators, which means execution quality is increasingly a differentiator.
What this means for underwriting
A few practical implications for how we underwrite PJM storage projects at AM Energy:
- RegD remains the default for fast storage. Storage assets without compelling reasons to enroll in RegA should default to RegD enrollment.
- Performance score sensitivity matters. Project pro formas should run sensitivities at 0.92, 0.95, and 0.98 performance scores — not just a single "0.97 for all years" assumption.
- Regulation can't be the whole thesis. Even strong regulation revenue doesn't justify a standalone storage project on its own. Regulation should be one revenue line in a stacked underwriting model that also includes capacity, energy arbitrage, and reserves.
- Regulation capacity competes with other dispatch. Every MW of regulation capability committed reduces the MW available for other revenue streams. Optimal dispatch models should optimize across the full revenue stack, not maximize regulation in isolation.
Where regulation is heading
The PJM regulation market is a finite-size market — total cleared capability across all hours is bounded by PJM's regulation requirement. As more fast resources qualify, the per-MW revenue available to any individual resource will continue to compress. We expect:
- Continued downward pressure on regulation clearing prices in well-saturated zones
- Higher relative value for regulation in constrained or under-saturated areas
- Increasing emphasis on combined regulation + reserves stacks, including new fast-response reserve products that PJM and other ISOs are exploring
None of this changes the fundamental case for utility-scale battery storage. It just reinforces that storage projects need to be underwritten across the full revenue stack — energy, capacity, regulation, and reserves — with realistic assumptions about how each market evolves.