A grid-scale battery, often called a BESS (battery energy storage system), makes money by buying power cheap and selling it dear. It also earns by being ready. It can steady the grid's frequency within a second, stand by as reserve, or be there on a cold winter evening. In Great Britain each of these is a separate market with its own rules and prices.
This article is for anyone joining a battery owner, an optimiser or a trading desk, or building their systems. It explains each market, how a battery chooses between them, and who takes the risk. It ends with one battery's summer day worked through.
What a battery sells
A battery has three things to sell:
- energy: megawatt-hours (MWh) bought at one time and sold at another;
- speed: it can go from charging to full output in under a second, faster than any power station;
- availability: the promise to be ready, whether or not it is called on.
The UK had 7.5 GW of grid-scale batteries at the end of 2025, by the Government's count, with a record 2.3 GW added that year. Modo Energy counted 8.5 GW in Great Britain at the end of September 2026. Most can run at full power for one or two hours. Over 2025, British batteries took in 2.7 TWh and gave back 2.3 TWh: about 85% of what goes in comes back out.
Trading the daily spread
The simplest income is arbitrage: charge when power is cheap, discharge when it is dear. Prices are lowest at night or, in summer, at midday when solar floods the grid. They are highest in the early evening. The gap between the day's dearest and cheapest hours is the spread.
Our analysis of the British day-ahead auction found the daily spread averaged £77.65 per MWh in 2026, up to 15 October. In September 2026 it reached £121.70, against £76.30 a year earlier, mostly because gas prices rose. Britain's batteries: where did the money come from? takes that spread apart.
A battery trades in the day-ahead auctions and then in the intraday market, as forecasts change. How power trading works explains those markets, and algorithmic power trading shows how a battery's bidding is automated. Two costs come off every cycle: the energy lost in the battery, and the wear on its cells.
The Balancing Mechanism
From one hour before each half-hour, called gate closure, NESO, the system operator, balances the grid itself. It does this in the Balancing Mechanism. A battery registered there sends two kinds of price:
- an offer: the price at which it will export more than planned;
- a bid: the price at which it will export less, or charge more.
When NESO accepts an offer, it pays the battery the offer price, and the battery exports. When it accepts a bid, the battery pays the bid price, or is paid if the bid price is negative. It then holds the energy to sell later. Batteries suit the Balancing Mechanism because they respond in seconds.
The catch is that NESO does not always take the cheapest action. When it passes over a cheaper offer, the offer is said to be skipped. The Government reports that batteries' skip rate fell from 49% in the first half of 2025 to 38% in the first half of 2026. NESO's own data shows it back at 47% in October 2026. A battery cannot count on its offers being taken.
Frequency response and reserve
The grid's frequency has to stay close to 50 Hz. If a power station trips, frequency falls within seconds, and something must make up the loss at once. Batteries are ideal for this. NESO buys these services in a daily auction, in six four-hour blocks, and everyone accepted in a block is paid the same price.
| Service | What it does | Full response within | Our calculation of the 2026 price, to date |
|---|---|---|---|
| Dynamic Containment | Rescues the grid after a sudden fault | 1 second | about £3.40 per MW per hour |
| Dynamic Moderation | Holds frequency steady in volatile periods | 1 second | about £7.60 per MW per hour |
| Dynamic Regulation | Corrects small, continuous drifts | 10 seconds | about £17.20 per MW per hour |
These prices are our own volume-weighted averages of NESO's published auction results. They are for the "low" direction of each service, where the battery exports when frequency falls.
NESO also buys reserve: power on standby that it can call on. Balancing Reserve started in March 2024. Quick Reserve, at full output within a minute, held its first auction in December 2024. Slow Reserve, within 15 minutes, held its first auction in March 2026 and replaced the old Short Term Operating Reserve from April. In 2026 they have cleared at about £3.70 to £4.25 per MW per hour.
The price is paid for being ready. But a battery holding 50 MW for frequency response in a block cannot use that 50 MW to trade. It must also keep enough energy to respond. That trade-off is the heart of running a battery.
The Capacity Market
The Capacity Market pays power stations and batteries a yearly fee to be available when the grid is tight, mostly on winter evenings. Its auctions are held four years ahead and one year ahead. Each battery's capacity is discounted for how long it can run, so a one-hour battery earns much less per MW than a power station.
The prices have fallen sharply. The four-year-ahead auction for 2029/30, held in March 2026, cleared at £27.10 per kW a year, after £60 to £65 in earlier rounds. The one-year-ahead auction for 2026/27 cleared at £5.00. Batteries won 1.2 GW of the former, after the discount.
Choosing between the markets
The markets do not simply add up. The same megawatt cannot be sold twice in the same half-hour, and the same megawatt-hour can only be sold once. So a battery owner, or the optimiser working for it, decides every day where each megawatt and each megawatt-hour earns most:
- Hold power for frequency response, and lose the chance to trade it in that block.
- Trade it day-ahead, and lock in a price.
- Keep it back for the Balancing Mechanism, and hope for a higher price, at the risk of a skip.
- Cycle less, because every cycle wears the battery, and its warranty usually limits how many cycles it may do.
Wear has a cost. A simple way to price it is to divide the cost of the cells by the cycles they will last. That might give a few pounds per MWh. A cycle that earns less than its wear is not worth doing.
Routes to market: who takes the risk
A battery owner also decides who trades the battery and who keeps the risk. The choices run from most risk to least:
- Merchant: the owner trades it, or pays a route-to-market provider a fee to send its trades and notices. The owner keeps every pound earned and every pound lost.
- Optimiser: a specialist trades the battery for a share of the revenue. The owner still takes the market risk.
- Floor: a trading house guarantees a minimum yearly revenue and takes a share of anything above it.
- Toll: a trading house pays a fixed fee for the battery's capacity, takes control of it, and keeps the trading result.
Both of the last two are now common in Britain. In July 2025 Gresham House Energy Storage Fund agreed floors on 789 MW, guaranteeing at least £35 million a year once in force. In January 2026 Drax agreed a ten-year toll on Fidra Energy's 250 MW battery at West Burton, at a fixed yearly fee rising with inflation. Floors dominated the contracts disclosed in 2025; in early 2026 the market began turning towards tolls.
Lenders like floors and tolls, because they make the income predictable. The owner gives up some upside in return.
Two current issues
Skip rates. Owners have long complained that NESO skips batteries in the Balancing Mechanism. NESO has built a new platform for sending instructions to batteries and small units in bulk. The skip rate fell in the first half of 2026, but has risen again since.
Re-trading behind constraints. When a grid bottleneck stops power flowing south, NESO may accept a bid from a battery behind it to cut its export. The battery can then sell that energy again for a later half-hour, and NESO may have to accept another bid. Ofgem, the regulator, found in June 2026 that this re-trading is not prohibited, though a false notice to NESO would breach the rules. NESO estimates it cost up to £97.9 million in 2025, and pumped hydro storage, not batteries, caused most of it. NESO is consulting on changes until 26 November 2026, and expects to make a proposal in early 2027.
Worked example: Northfield Battery's summer day
Northfield is an invented trading firm. Northfield Battery is a 100 MW battery that stores up to 200 MWh: two hours at full power. Of every 100 MWh it buys, it can sell back 86. It trades merchant: a route-to-market provider sends its trades and notices for a fee. Every price below is made up for the example. The day is Thursday 16 July 2026, and times are UK summer time (BST).
The plan
Sunshine pushes prices down at midday, and the evening peak comes after sunset. So the battery charges at midday, holds its charge through the afternoon, and sells in the evening.
| Time | What the battery does | Working | £ |
|---|---|---|---|
| 11:00 to 13:00 | Buys 200 MWh day-ahead at an average £15 | 200 x £15 | -3,000 |
| 15:00 to 19:00 | Holds 50 MW for Dynamic Containment at £3.40 per MW per hour | 50 x 4 x £3.40 | +680 |
| 19:00 to 20:30 | Sells 150 MWh day-ahead at £104, £112 and £108 | 50 x (104 + 112 + 108) | +16,200 |
| 20:30 to 21:00 | Its last 22 MWh are accepted in the Balancing Mechanism at £150 | 22 x £150 | +3,300 |
| All day | Wear, at an example £5 per MWh sold | 172 x £5 | -860 |
| All day | The provider's fee, at an example £1 per MWh traded | 372 x £1 | -372 |
| The day | +15,948 |
The choices inside it
- Why hold Dynamic Containment from 15:00 to 19:00? The battery is full and waiting for the evening. Being ready for a fault costs it almost nothing in that block. It keeps enough energy to export 50 MW for 15 minutes if called.
- Why keep 22 MWh back? The battery has only 172 MWh to sell after its losses. Northfield bet the Balancing Mechanism would pay more than the last day-ahead half-hour. Had NESO skipped the offer, Northfield would have sold the energy intraday, perhaps for less.
- Why not cycle twice? A second cycle would earn a smaller spread than its wear.
If Northfield had tolled it
Suppose Northfield had instead tolled the battery to a trading house for a made-up £60,000 per MW a year. That pays £16,438 a day, whatever the market does. On this good summer day the merchant battery earned a little less than the toll. On a day with a narrow spread it would earn far less. The toll swaps that uncertainty for a fixed fee.
What an optimiser's systems have to do
Running a battery well is mostly a software problem. An optimiser's systems have to:
- track the state of charge every few seconds, and know what energy is promised to each market;
- forecast prices for every half-hour, and the chance of each Balancing Mechanism offer being taken;
- bid into each auction on time: the frequency and reserve auction, the day-ahead auctions, and intraday trading up to gate closure;
- keep the promises consistent: never sell the same megawatt to two markets, and keep energy for every service held;
- count the cycles against the warranty, and price the wear;
- send physical notices to NESO and react to its instructions within seconds;
- reconcile every payment from NESO, the exchanges and the Capacity Market against what the battery actually did.
Algorithmic power trading shows the loop such a system runs, and the checks it needs.
Building or replacing this on your own desk? It is the work A115 does. See the systems work
What the full guide adds
This article is the short version. The full guide, Revenue Markets: Where Batteries Earn in GB and Europe, is part of the BESS Economics course:
- The four families of income and the market access a battery needs for each;
- Wholesale and balancing in depth, including how skips happen;
- The frequency services, and why their prices collapse as more batteries join;
- The Capacity Market, with a worked discount for duration;
- A revenue map for one battery and one evening, service by service;
- Other European markets, for comparison.
Two further guides in the same course go deeper. Revenue Stacking explains why the stack is not additive, and Route to Market covers floors, tolls and what lenders want.
The full guide is in the community on Skool, with the rest of the series, and it is updated regularly as the markets change. If you want to work on a trading desk, read it to understand the subject properly. To join, start at the Energy Trading Fundamentals course page.
Sources
- DESNZ, Grid-scale battery storage statistics (Energy Trends, June 2026): capacity, output and efficiency.
- Modo Energy, GB BESS buildout Q3 2026 (October 2026): 8.5 GW at the end of September.
- DESNZ, Clean Flexibility Roadmap: July 2026 update: battery skip rates. NESO, skip rates.
- NESO, Dynamic services (DC, DM, DR), Enduring Auction Capability, and the EAC auction results on the NESO Data Portal, from which we calculated the 2026 prices.
- NESO, Quick Reserve, Slow Reserve and Balancing Reserve.
- Ofgem, Annual report on the operation of the Capacity Market in 2025-26 (September 2026): auction prices and battery volumes.
- Ofgem, Repetitive re-trading by electricity storage in transmission constraint periods (June 2026), and NESO, Storage behind constraints consultation (October 2026).
- Gresham House Energy Storage Fund, floor agreements on 789 MW (July 2025); Drax, announcement of its toll with Fidra Energy (30 January 2026); Pexapark, GB BESS shift to tolling (March 2026).