Irish batteries discharge about once a day. British batteries discharge about four times a day
The same 54 days of meter readings for every grid battery on the island of Ireland and every one in the balancing mechanism in Britain. The Irish fleet discharges into the morning and evening peaks and does little else. 13 of its 23 working batteries charge at less than half the rate they discharge, and nine have never run above 57% of their registered power. Each of those is a trace of the payments the fleet was built for.
Read from SEMO's daily meter data and Elexon's B1610, trade dates 12 August to 4 October 2026, first settlement reading. 23 Irish batteries (708 MW) and 137 British units (5,812 MW) that ran in the window. Measured, not modelled.
Both fleets discharge into the morning and evening peaks. Only the British fleet works through the night and the middle of the day
The chart below is the average day of a battery in each country. Every half hour's output is divided by that battery's own peak, so a 20 MW unit and a 100 MW unit count the same, and the two fleets can sit on one scale. Discharging rises above the line, charging falls below it.
The two shapes agree on the peaks. Both fleets empty into the evening peak between 17:00 and 21:00, and the top of the Irish curve at 19:00 is a little below Britain's. Both discharge again in the morning around 07:00, the British fleet at more than twice the Irish rate. Between the peaks they part company. The British fleet keeps a few per cent of its power moving through the night and the middle of the day, and fills in two sittings, one overnight and one at midday. The Irish fleet is close to silent from midnight until 06:00 and again from 09:00 until 17:00, and it fills mainly at midday. On an island with Ireland's wind, the empty night is the odd one.
Average charging and discharging by half hour, local time, as a share of each battery's own peak power, 12 August to 4 October 2026. Each battery day counts equally. Source: SEMO Daily Meter Data and Elexon B1610, read by Vitreous Labs.
Counting the separate discharges makes the same point. A British battery discharges 3.8 times a day on average, an Irish one 1.25 times. We read the British pattern as the balancing mechanism and frequency response at work, a battery dispatched for a few half hours at a time whenever the system needs it. Ireland's market is organised differently, and the meter shows an Irish battery earning on the evening spread and otherwise waiting.
13 of Ireland's 23 working batteries charge at less than half the rate they discharge
A battery that charges slowly is not ready for the next evening peak, so it can trade once a day whatever prices do. That is what this section measures. Each dot below is one battery. Across is the fastest it discharged in any half hour of the 54 days, up is the fastest it charged. A battery that fills as fast as it empties sits on the diagonal. Nearly every British unit does. Thirteen of the 23 working Irish units sit well below it, and the biggest gap is at ESB's Aghada site, where two 75 MW units never charged faster than 16 MW.
Fastest half hour of charging against fastest half hour of discharging per battery, in MW, 12 August to 4 October 2026. Source: SEMO Daily Meter Data and Elexon B1610, read by Vitreous Labs. Hover a dot for its name.
This caps the cycling before prices get a say. A 150 MWh battery that can only take 16 MW needs nine hours to refill. Why Irish connections are built this way is not in the meter. The likeliest reading is that the connection agreement carries a separate maximum import capacity, set low when the battery was expected to live on availability payments rather than trading. We have not seen the agreements and say so.
Nine Irish batteries have never produced more than 57% of their registered power in any half hour
A battery that never uses more than half its power is being held back for something other than selling energy. This pattern is in the Irish fleet alone. The chart puts each battery's highest half hour output beside the capacity SEMO registers for it. Every energy battery, the two hour and four hour units, reaches its registered power. Nine of the older units, most of them registered at 50 MW and built for the fast frequency services, stop between 46% and 57% and did not pass it on any of the 54 days. A burst at full power for a few minutes would average well under half in a half hour, so the shape fits a battery used for short services and never for a sustained discharge. Whether it is a cap in the service contract or in the connection agreement is the question to put to the owners before it is said as fact.
Highest average output in any half hour, 12 August to 4 October 2026, as a share of SEMO's registered capacity, July 2026 report. Batteries under 5 MW at peak or active fewer than 10 days are left out. Source: SEMO Daily Meter Data and Registered Capacity report, read by Vitreous Labs.
The Irish fleet is two fleets: nine energy traders and fourteen short reserve batteries
The three measurements sort the same batteries into the same two groups. Nine batteries behave like British traders: Poolbeg, Raghra, the three Aghada units, South Wall, Kylemore, Cloncreen and Ballykilleen. They hold 1.7 to 4.1 hours of energy at full power, return 78% to 87% of what they take in, and do the one evening cycle a day the market gives them. The other fourteen are half hour and one hour units that move very little energy and return between 32% and 70% of it, because standing losses dominate a small throughput. Five more exported almost nothing in 54 days and are left out of the averages.
| Battery | Registered MW | Peak out MW | Peak in MW | Hours at peak | Exported MWh | Out ÷ in | Cycles a day | Active days |
|---|---|---|---|---|---|---|---|---|
| Poolbeg BESS | 75 | 74.5 | 74.6 | 1.8 | 5,794 | 81% | 0.78 | 51 |
| Raghra 220KV | 63.2 | 62.8 | 62.2 | 2.3 | 5,304 | 84% | 0.67 | 46 |
| Aghada BESS_02B | 75 | 74.3 | 15.6 | 1.9 | 3,085 | 78% | 0.41 | 46 |
| Aghada BESS_02A | 75 | 74.8 | 15.7 | 1.7 | 3,067 | 78% | 0.46 | 46 |
| Cloncreen BESS | 25 | 25.1 | 15.3 | 2.5 | 2,424 | 86% | 0.72 | 38 |
| Ballykilleen BESS | 20 | 19.9 | 20.7 | 4.1 | 2,266 | 87% | 0.51 | 46 |
| Southwall BESS | 30 | 29.7 | 28.6 | 1.9 | 2,252 | 81% | 0.76 | 49 |
| Kylemore BESS | 30 | 29.7 | 28.7 | 1.8 | 2,218 | 78% | 0.77 | 51 |
| Aghada BESS | 19 | 18.7 | 17.1 | 1.8 | 1,389 | 81% | 0.75 | 52 |
| Kells Battery Storage NI | 50 | 25 | 25 | 1.0 | 935 | 62% | 0.70 | 47 |
| Castlereagh BESS NI | 50 | 27.4 | 22.8 | 0.9 | 872 | 55% | 0.66 | 50 |
| Porterstown BESS | 30 | 29 | 4.6 | 0.6 | 705 | 62% | 0.73 | 44 |
| Lumcloon ESS A-1 | 50 | 26.5 | 7.2 | 0.7 | 675 | 64% | 0.66 | 49 |
| Shannonbridge BESS (A1) | 50 | 28.6 | 7.2 | 0.6 | 674 | 63% | 0.70 | 48 |
| Shannonbridge BESS (A2) | 50 | 26.6 | 7.2 | 0.7 | 669 | 69% | 0.66 | 48 |
| Lumcloon ESS A-2 | 50 | 26.5 | 7.5 | 0.7 | 669 | 70% | 0.65 | 49 |
| Mullavilly Storage NI | 50 | 27.6 | 29.1 | 0.6 | 533 | 51% | 0.60 | 48 |
| Kelwin Power Plant Phase 2 | 26.6 | 18.2 | 3.9 | 0.5 | 398 | 39% | 0.74 | 50 |
| Bannbridge Main NI | 12.2 | 12 | 2.4 | 1.3 | 319 | 109% | 0.38 | 28 |
| Lisdrumdoagh BESS | 60 | 30.5 | 13.6 | 0.5 | 197 | 32% | 0.24 | 34 |
| Killala Battery Storage | 10.8 | 9.2 | 4 | 0.9 | 179 | 60% | 0.42 | 31 |
| (unnamed, GU_404510) | 8.5 | 6.5 | 2 | 0.8 | 143 | 66% | 0.54 | 45 |
| BEEKTOBS | 11 | 5.1 | 1.7 | 0.7 | 122 | 40% | 0.62 | 54 |
| Gorman BESS | 50 | 9.9 | 5.1 | 1.8 | 20 | 2% | 0.02 | 2 |
| Avonbeg Storage | 16 | 7.8 | 3 | 0.7 | 7 | 5% | 0.02 | 1 |
| Drumkee Storage NI | 50 | 5 | 4.3 | 1.1 | 6 | 1% | 0.02 | 1 |
| Gorey BESS | 9 | 1.8 | 2 | 0.5 | 2 | 2% | 0.03 | 4 |
| Golagh Battery | 3 | 0 | 0 | 0 | 0.00 | 0 |
Every battery SEMO files as storage, 12 August to 4 October 2026. Hours at peak is the longest uninterrupted discharge divided by the battery's peak power. Out ÷ in is energy exported over energy imported; for a battery that hardly cycles it is mostly standing load, not efficiency. Bannbridge exports more than it imports, which is a metering question, not a result. NI marks Northern Ireland. Source: SEMO Daily Meter Data and Registered Capacity, read by Vitreous Labs.
Why: Ireland paid its batteries to be ready, Britain pays them to move energy
Ireland's first wave of batteries was built for DS3, EirGrid's programme for running the grid on high levels of wind. DS3 pays a reserve provider for every hour it is available, with multipliers for speed and for scarcity. It pays for being ready, not for energy delivered. A battery built to earn that way wants to be large in power, small in energy, and parked part charged so it can respond either way. The nine half power units are what that money bought.
Britain pays the other way round. A British battery earns from the wholesale spread, from the balancing mechanism, from frequency response bought in daily auctions and from the capacity market, and all of them reward a two hour unit that charges and discharges at full power several times a day. The meters show a fleet shaped by that.
The Irish side is changing. DS3 was meant to end this year and has been extended until September 2027 at the latest, with a daily auction for reserves, the Day-Ahead System Services Auction, due in May 2027. The two hour and four hour batteries already on the island behave like British ones and are the shape of what comes next. The half hour units are the ones with a question over them.
What this does not show
- 54 days from late summer into autumn, so nothing here is a seasonal claim. Winter will look different in both countries. The window is short because the public record is: SEMO deletes its daily meter report after about two months, and Irish batteries are not reported to the European transparency platform at all. Vitreous Labs has saved every day since 12 August 2026 and keeps saving, so this comparison will be rerun as the record grows.
- SEMO's figure is the first reading, published the day after trading, and can be revised. Elexon's can be revised for months.
- Cloncreen, Kelwin 2 and Killala sit beside wind farms and some charging may come from the turbines behind the grid meter. Their figures do not suggest it, but it cannot be ruled out from the meter.
- Why the five idle batteries were idle (Gorman, Avonbeg, Drumkee, Gorey, Golagh) is not known. A flat zero can be an outage, a contract ending, or a meter that is not reporting.
- The causes offered for the slow charging and the half power cap are readings of the shape, not documents we have seen.
Sources. SEMO, Daily Meter Data (Metered Generation by Unit), every half hour, saved daily by Vitreous Labs since 12 August 2026. SEMO, Registered Capacity report, July 2026. Elexon, B1610 actual generation output per generation unit. EirGrid, DS3 System Services tariffs to FASS, the Gap, recommendations paper, June 2025.