A familiar battery pitch deck has one especially attractive slide. It shows ancillary-service prices, multiplies them by the project's megawatts, and carries the result across the forecast period.
The logic feels solid. The grid needs fast response. Batteries are exceptionally good at providing it. Prices are attractive, cycling is relatively shallow, and the asset can earn money without waiting for a large energy-price spread.
Then more batteries arrive.
The grid may benefit from every one of them, but the ancillary-service market does not automatically buy more simply because more supply is available. It buys the quantity needed to manage frequency, contingencies, and short-term uncertainty. Once qualified supply grows faster than that requirement, the economics change quickly.
The service still matters. The line in the pitch deck does not necessarily survive.
Saturation is a ratio, not a date
An ancillary-service market is becoming saturated when qualified, deliverable supply is persistently large relative to the quantity the system operator procures.
I would not call a market saturated because one quarterly price chart turns down. Saturation is not the day when one more battery connects. It is not proof that the grid has too much storage. It does not mean ancillary services have become worthless.
It means the market has moved from scarcity of capable providers to competition among capable providers.
The denominator is the requirement. The numerator is the supply that can actually clear after location, telemetry, response speed, state of charge, duration, and market rules are considered.
If a system procures 500 MW of regulation and 550 MW can provide it, the last qualified megawatt may be valuable. If 2,000 MW can provide the same 500 MW, three quarters of the fleet must find another job in that interval.
This is why installed battery capacity alone tells you very little. What matters is eligible supply divided by procured demand, hour by hour and product by product.
Why ancillary services saturate early
Ancillary services can be high-value markets, but they are usually shallow markets.
The U.S. Energy Information Administration describes them as high-value but low-volume in its battery-storage analysis. That is exactly the combination that attracts early investment and then compresses quickly.
Batteries accelerate the process for three reasons.
They are technically well matched to the product. A battery can change output rapidly, follow an automatic control signal, and move both upward and downward when it has sufficient energy and headroom.
A large share of the fleet can qualify. Regulation is not limited to a handful of unusually configured projects. Once market access, controls, and telemetry become standardized, many batteries can chase the same requirement.
Their short-run offer price can be low. When degradation, activation energy, and opportunity cost are modest, a battery may rationally offer ancillary capacity near zero. In a pay-as-cleared market, several low offers can push the clearing point down for everyone.
Meanwhile, the amount procured is driven by system risk. Requirements can grow as load, renewable variability, or contingency exposure changes, but they rarely grow one-for-one with battery installations.
That mismatch is the basic engine of saturation.
Price compression arrives before the market disappears
Imagine an illustrative uniform-price market that procures 500 MW of a reserve product each hour. To keep the mechanics visible, assume one capacity payment and ignore mileage, performance multipliers, activation energy, and uplift.
At first, only 550 MW of qualified supply is competing. The marginal accepted offer sets a clearing price of $18/MW-hour.
1. Total market payment: 500 MW x $18/MW-hour = $9,000 per hour.
2. Marketwide payment pool divided by competing qualified supply: $9,000 / 550 MW = about $16.36 per hour.
Now 600 MW of low-cost battery supply enters. The requirement stays at 500 MW, qualified supply rises to 1,150 MW, and the marginal offer falls to $6/MW-hour.
3. Total market payment: 500 MW x $6/MW-hour = $3,000 per hour.
4. Marketwide payment pool divided by competing qualified supply: $3,000 / 1,150 MW = about $2.61 per hour.
The clearing price fell by two thirds. The payment pool relative to competing qualified supply fell by roughly 84 percent. That second figure is not a forecast of any individual project's earnings; merit order and operating availability determine who actually clears.
No one cancelled the service. The system operator still procured 500 MW. The market simply stopped paying a scarcity premium for a capability that had become abundant.
This example is deliberately simple. Actual markets use different offer rules, performance adjustments, co-optimization, locational constraints, and settlement formulas. The point survives those details: when supply grows faster than the requirement, both price and award frequency come under pressure.
California shows the transition in real time
California is a useful case because battery growth is no longer theoretical.
The California ISO's 2024 Special Report on Battery Storage says active battery capacity in its balancing area grew from about 500 MW in 2020 to 13,000 MW by December 2024. Batteries provided nearly 84 percent of regulation up and regulation down requirements in 2024.
But the same report says batteries consistently offered many times more regulation than the market required. Battery capacity grew faster than ancillary-service requirements, so the share of the battery fleet scheduled for those services declined.
For upward regulation, spinning reserve, and non-spinning reserve combined, the portion of total battery capacity scheduled fell from 12 percent in 2023 to 9 percent in 2024.
Where did the batteries go? Increasingly, into energy.
CAISO reported that batteries earned nearly 82 percent of their market revenue from energy in 2024. Average battery output during the evening peak also rose sharply, reaching about 5,700 MW in hour-ending 19, compared with 2,700 MW in 2023.
This is the part a revenue headline can miss. The battery fleet does not stop working. It migrates toward a deeper value pool.
Deeper does not mean bottomless. As storage penetration grows, batteries can eventually compress energy spreads too. But the energy market is tied to the full system dispatch, while a regulation requirement may be only a narrow slice of system capacity. The transition buys a project more room, not immunity from competition.
PJM shows why market design matters too
PJM offers a different warning.
In its 2024 State of the Market Report, PJM's independent market monitor said there was operational evidence that the RegD market was saturated. It also argued that flaws in the regulation design had encouraged over-procurement and sent distorted investment signals.
There is a trap here. A market can look lucrative because the system genuinely lacks a service, because the rules overvalue a particular response characteristic, or because both are true at once.
Historical prices can look like facts about physics. Sometimes they are partly facts about market design, and market design can be rewritten.
PJM implemented the first phase of a regulation-market redesign in October 2025, including a move toward a single bidirectional signal and 30-minute intervals. PJM said the changes were intended to reduce over-procurement and improve efficiency, with another phase planned for 2026. Its own review of the changes makes the lesson clear: when a product becomes operationally or economically distorted, rules can change faster than a project's capital structure.
Not every ancillary-service product saturates together
“The ancillary-services market” is not one pool.
Regulation, spinning or synchronized reserve, non-spinning reserve, contingency products, voltage support, black start, and flexible ramping products solve different system problems. They have different procurement volumes, response times, duration requirements, and qualification rules.
Regulation often feels saturation first because batteries fit it so well and the requirement is relatively small. A longer-duration contingency reserve may remain less crowded. A local reliability product can stay scarce even while a systemwide product is oversupplied. New products may also emerge as the grid's risk profile changes.
So the right question is not, “Are ancillary services saturated?”
It is:
Which product, in which zone, during which hours, under which rules, is short of qualified supply?
That is a much less convenient question. It is also the one the revenue model needs to answer.
What saturation does to a battery business case
The first casualty is the base case built from old ancillary-service prices.
A backcast can show what a battery would have earned before thousands of competing megawatts arrived. It cannot tell you what the next battery earns after it joins the competition. That sounds obvious, yet it is one of the easiest errors to hide inside a polished model.
As ancillary revenue compresses, five things usually happen.
Revenue shifts from availability to dispatch. The project earns less for standing ready and more by moving energy. That increases exposure to charge prices, discharge prices, efficiency, and forecasting.
Duration matters more. A one-hour battery may be excellent at regulation. A four-hour battery has more freedom to move into evening energy, capacity, or longer reserve products when regulation prices weaken. Extra megawatt-hours do not create value by themselves; they create options only where another market or contract pays for them.
Cycling and degradation assumptions change. Shallow regulation movements and deep arbitrage cycles do not consume the asset in the same way. A revenue pivot requires a new degradation case, not just a new price row.
Optimization becomes commercially important. When one easy product no longer carries the economics, value comes from switching among several imperfect opportunities without violating state-of-charge or contract constraints.
Contract quality becomes more visible. A toll or capacity agreement can protect the owner from merchant price compression, but only if the payment and performance obligations are genuinely bankable. A vague “revenue stack” does not replace contracted cash flow.
The asset may still be attractive. It is simply a different asset than the one described by the early-market backcast.
Saturation does not mean the grid has too many batteries
This is the most important distinction in the article.
A grid can have an oversupplied regulation market and still need much more storage for energy shifting, resource adequacy, congestion relief, renewable integration, resilience, or future load growth.
Ancillary-service saturation says that one narrow product has enough qualified megawatts at prevailing requirements and rules. It says nothing conclusive about the system's multi-hour energy needs.
In fact, cheap ancillary services can be a public success. Customers receive a reliability service at lower cost. The problem belongs to the investment case that assumed the early scarcity rent would last forever.
Lower prices are not evidence that batteries failed.
They may be evidence that batteries solved the first problem they were paid to solve.
Saturation can reverse - but not on your schedule
A saturated market is not necessarily saturated forever.
Load can grow. Renewable variability can increase procurement needs. Old providers can retire. Qualification rules can tighten. A redesigned product can value duration, location, or sustained response in a way that reduces the truly substitutable supply pool.
Scarcity events can also produce occasional price spikes even when ordinary hours remain oversupplied. Those spikes preserve optionality, but they are a fragile foundation for fixed debt service.
The sensible forecast is therefore not a straight line to zero. It is a set of regimes: ordinary oversupply, occasional scarcity, possible requirement growth, and possible rule changes. The project should survive the first regime before investors give it credit for the others.
How to underwrite the market after the premium fades
A credible investment case should treat ancillary services as a competitive market, not an annuity.
The strongest question is brutally simple:
If this revenue line falls by 70 percent, what job does the battery do instead?
If the answer is “the same job at the same volume,” the model has not absorbed the meaning of saturation.
The market buys a quantity, not a story
Ancillary-service markets reward batteries for being fast, accurate, and available. They do not promise to reward every fast battery equally, or forever.
Early projects can earn scarcity rents because few assets can provide the product. Successful technology attracts capital. Capital adds qualified supply. Prices compress. The fleet moves toward energy, capacity, local reliability, or contracts.
That sequence is not an accident. It is how a competitive market is supposed to work.
The mistake is not earning ancillary-service revenue.
The mistake is financing a 15-year asset as though a shallow market will remain undersupplied for 15 years.