Frequency regulation is usually the first market a new battery fleet conquers, the first line item in its early revenue stack, and the first paycheck to shrink. This article is about how that service actually pays — the two payment shapes, why batteries beat everything else at earning them, what the duty does to the cells, and why the market that made early batteries rich keeps doing the same trick to each new region and then stopping.

If you want the one-paragraph version: frequency regulation pays a battery to continuously follow a dispatch signal that corrects the grid’s second-by-second supply-demand imbalance. It pays twice — once for being available, once for how well and how far you move. Batteries follow the signal almost perfectly, so they take over the market, and because the market is shallow, taking it over crushes the price. Everything else is detail, but the detail is where projects get underwritten well or badly.

What you’re actually paid to do

Grid frequency sits at 50 or 60 Hz only when generation exactly matches load. It never exactly does, so system operators buy a standing correction service: units that hold headroom and move their output up or down on command, second by second, to close the gap. Depending on where you operate, the product is called Regulation Up/Down in the US ISOs, FCR and aFRR in Europe, or regulation FCAS in Australia — different acronyms, same job: track the signal.

The signal wanders about zero and roughly cancels over time, so what binds is power rating, response accuracy and round-trip efficiency — plus enough state-of-charge headroom in both directions to keep following it accurately — rather than deep energy capacity.
one new command — typically every 2 to 4 secondsregulation command0dischargechargetime — seconds to minutesthe two shaded totals, drawn to one scaledischarge — injecting powercharge — absorbing itclose, not equal — the signal is not perfectly neutralAssets are commonly sized at about a 15-minute to 1-hour energy-to-power ratio — set bymarket rules, not chemistry.Illustrative signal: no market data stands behind this shape, and no value can be readoff either axis. What it asserts is that the command swings both ways and that the twoshaded totals roughly cancel; the size and direction of the gap between them areillustrative too, while the step width and the sizing range below are measured figures,not the trace’s.

The payment has two shapes, and the split matters more than the names:

  • Capability (availability). A price on the MW of headroom you commit — you’re paid to stand ready whether or not the signal moves you far. This is the reserve-like half of the product.
  • Performance (mileage). A price on the movement itself — how far you actually travel following the signal, weighted by how accurately you track it. In the US this owes its existence to FERC Order 755, which forced markets to pay for performance rather than treating a sluggish steam unit and a millisecond-fast battery as the same megawatt.

Accuracy scoring is where batteries quietly win: the performance score multiplies effective revenue, so an inverter-based plant tracking the signal at near-perfect fidelity out-earns a thermal unit offering the same MW — before prices even move.

Why batteries own this market

Three physical advantages, none of which a competitor can buy back:

  1. Speed. A PCS reaches a commanded setpoint in a fraction of a second. The generation it displaced took tens of seconds to minutes.
  2. Precision. The inverter produces exactly the commanded power, so signal-tracking scores sit near the top of what the scoring formula allows.
  3. Symmetry. From mid-SOC a battery regulates up and down with the same hardware, which lets it sell both directions of the product at once where the market allows it.

There’s a fourth, economic advantage: regulation is a power product, not an energy product. The signal is roughly energy-neutral over time — you’re jittering around a setpoint, not discharging for hours — so a short-duration battery can sell its full power rating into regulation without owning the energy to sustain it. That’s why young merchant fleets skewed short-duration and ancillary-heavy, and why the later shift toward capacity and arbitrage revenue pushed the same market toward 4-hour systems.

The proof is in every market that let batteries in. PJM’s fast RegD signal became a battery playground within a few years of Order 755 (and PJM later had to rework the signal and its rules when the fast signal turned out not to be energy-neutral in stressed hours — a reminder that market design iterates). Great Britain’s Dynamic Containment launched in 2020 at prices that briefly made batteries some of the fastest-payback grid assets anywhere; the DC/DM/DR suite then saturated within a few years. ERCOT ran the same movie earlier and more completely: early storage earned most of its money in ancillary services, fleet growth crushed clearing prices, and by 2024–2025 energy arbitrage had become the dominant earner.

The saturation cycle — the paycheck’s half-life

The pattern repeats because the cause is structural: regulation demand is small and roughly fixed — a grid needs on the order of a percent of peak load held as regulating capability, and that requirement doesn’t grow just because batteries showed up. Supply, meanwhile, arrives in hundred-megawatt blocks every quarter. A shallow market plus a steep supply curve equals a price cliff.

So the revenue trajectory of “battery in a young market” is predictable: frequency regulation and reserves pay anywhere from tens to over $100/kW-year while the fleet is small, then decay toward arbitrage levels within a few years as batteries flood in. ERCOT first; CAISO and Great Britain running it now; every market that opens to storage next is somewhere earlier on the same curve. For a deeper look at what takes over after saturation — arbitrage, capacity, and how the whole stack fits together — the revenue-streams deep dive covers the full portfolio.

The underwriting rule falls straight out of the physics: never underwrite on today’s regulation price. Model the fleet pipeline, assume the decay, and treat early-year ancillary revenue as a bonus that pays down capex — not a perpetuity. Lenders already do.

Here’s the distinction that trips up newcomers reading grid codes and market rules side by side: not all frequency-correcting behavior is a product you sell. A lot of it is an obligation you owe.

Frequency response — the autonomous droop reaction where the plant sees frequency deviate and adjusts power without any dispatch signal — is mandatory and largely uncompensated in several major markets. ERCOT requires primary frequency response from every generation resource and ESR, with governor droop of 5% or less and a deadband no wider than ±0.017 Hz, as a condition of connecting — the clauses are on our ERCOT grid-code page. Australia’s mandatory PFR regime binds all scheduled units including batteries the same way (Australia NER / AEMO). The battery does real work in both cases; nobody sends a check. (Australia is the partial exception since mid-2025: frequency performance payments now reward units whose response helps and charge the ones whose response hurts — an incentive layer bolted onto the mandate, not a market you bid into.)

Frequency regulation, by contrast, is procured: you bid, you clear, you follow the operator’s signal, you invoice. In Europe the paid product itself is droop-shaped — FCR and GB’s Dynamic Containment respond to measured frequency, not a dispatch signal — so the line runs between procurement and obligation, not between control schemes. When you model a project, keep the two in separate columns — the mandatory response consumes headroom and throughput that the paid products can’t sell, and a model that books the same megawatt twice will fail its first serious review.

What the duty does to the battery

Regulation cycles the battery shallowly but nearly continuously. Individually the cycles are small; summed over a year of signal-following, the energy throughput is real, and it belongs in the model twice:

  • Degradation. Cell wear tracks energy throughput and cycling conditions. A MW-year of regulation duty has a throughput bill measured in full-cycle equivalents, and the warranty’s cycling assumptions need to cover the duty profile you’re actually selling.
  • Efficiency. Every MWh that jitters through the battery pays the round-trip toll. Regulation revenue is quoted gross; the energy you buy back to hold SOC is a cost line.

State of charge is the operational headache. The signal is only roughly energy-neutral: hours of net-up or net-down movement drag SOC toward a rail, and a battery pinned at 5% or 95% SOC can no longer deliver both directions — which torpedoes the performance score that made it money in the first place. Fleets manage this with SOC bands, market energy transactions to re-center, and in some markets explicit rules about how storage restores its state of charge. It’s solvable, but it’s an operations discipline, not a footnote.

How to think about it in 2026

Frequency regulation is no longer the business case for storage in mature markets — it’s the appetizer. The durable way to hold it in your head: a small, fast-saturating market that batteries win on physics, worth real money early, structurally incapable of paying a whole fleet forever, and permanently useful as the highest-margin use of the first megawatts of any new fleet. Price it as a curve, keep it in a separate column from your grid-code obligations, charge the throughput to the model, and let the capacity and arbitrage lines carry the debt.

FAQ

How much does frequency regulation pay a battery? In a young, shallow market: anywhere from tens to over $100/kW-year. In a saturated one: a fraction of that, decaying toward arbitrage levels. The honest answer is a curve, not a number — never underwrite on today’s clearing price.

Is frequency response the same as frequency regulation? No. Frequency response is the plant’s autonomous droop reaction to a frequency deviation — much of it mandatory and unpaid under grid codes. Frequency regulation is a procured market service where you continuously follow the operator’s dispatch signal for money.

Why are batteries so good at frequency regulation? Speed, precision, and symmetry: an inverter reaches a commanded setpoint in a fraction of a second, tracks the signal almost exactly, and can move in both directions from mid-SOC. Scoring systems that pay for accuracy hand batteries the top of the merit order.

Does regulation duty wear the battery out? It cycles the battery lightly but almost continuously. The throughput is real and belongs in the model as a degradation and efficiency cost per MW of regulation sold — it rarely kills the business case, but ignoring it flatters the margin.

Revenue modeling, saturation dynamics, and how lenders stress-test ancillary assumptions — covered in depth in my Grid-Scale BESS: Complete Guide.