The Real Economics of Battery Storage in Ireland

01The Question

Here is the pitch you'll hear from every energy consultant in Dublin: Ireland is building enormous amounts of wind power. Wind is intermittent. When it blows hard at 3am and nobody wants the electricity, prices crash. When it dies at 6pm and everyone's cooking dinner, prices spike. A battery sits in a field, charges when electricity is cheap, and discharges when it's expensive. You pocket the difference. Simple.

The pitch is not wrong, exactly. It's just incomplete in ways that are worth about −EUR 21.4 million over twenty years.

We modeled a specific project: a 50 MW / 200 MWh lithium iron phosphate (LFP) battery, connected to the Irish transmission grid, with a commercial operation date of Q1 2028 and a 20-year economic life. This is not a small pilot — it's roughly the size of the Lumcloon Energy project, one of the largest operational batteries in Ireland. The question is whether it makes money.

This is the important part, pay attention: Under current Irish regulations, this project has an IRR of −3.5%. Not a low return. Not a marginal return. A negative return. You'd lose roughly two-thirds of the capital you put in. The total undiscounted free cash flow over 20 years is −EUR 21.4 million.

But here's why this report exists: a single regulatory change — exempting battery storage from a particular grid charge — would flip the IRR to roughly +8.7% to +10.8%, depending on timing. That's the difference between "burn your money" and "decent infrastructure return." Everything in this analysis is about understanding that gap and the probability of it closing.

02How Irish Electricity Pricing Actually Works

Ireland's wholesale electricity market is called the Single Electricity Market (SEM), and it's shared with Northern Ireland. It runs a day-ahead auction where generators bid to supply each half-hour of the next day, and a balancing market that settles differences in real time. The day-ahead price is what matters most for a battery, because that's where you plan your buy-low-sell-high strategy.

Here's what makes it weird: Ireland is a small, isolated island with a lot of wind and a lot of gas turbines, and not much else. In most hours, the marginal generator — the last one needed to meet demand — is a gas plant. This means the wholesale price is effectively set by the cost of burning gas in a combined-cycle turbine, which works out to roughly:

That EUR 101/MWh is the average price when gas is setting the margin. But averages hide the thing that matters for batteries: volatility. The daily price profile looks roughly like a valley with two peaks:

The trough happens around 3–5am, when demand is low and wind is often still blowing. Prices drop to EUR 50–80/MWh. The peak happens around 5–7pm (and sometimes a morning peak at 8–9am), when everyone gets home and the wind may or may not cooperate. Prices spike to EUR 130–200/MWh. On a good day for a battery, that's a spread of EUR 60–120/MWh.

Why does the spread exist? Because electricity demand fluctuates much more than supply can ramp smoothly. When wind drops off and gas plants have to fire up quickly, the price includes start-up costs, scarcity premia, and the general nervousness of grid operators who really don't want blackouts. When wind is roaring and demand is low, generators sometimes bid at zero or negative prices to avoid the cost of shutting down. The battery lives in this gap.

We measured this empirically. Using 74,969 hours of SEMOpx Day-Ahead prices from October 2018 to February 2026 (2,695 complete days), the median 4-hour spread (best any 4 cheap hours vs. best 4 expensive hours) is EUR 63.9/MWh, and the mean is EUR 79.8/MWh. The mean is 25% higher than the median because the distribution has fat tails — occasional days where the spread is EUR 200+ due to gas supply scares or wind droughts. On 30.8% of days, the net spread (after round-trip efficiency losses) is below EUR 40/MWh, barely enough to cover operating costs. The top 25% of days generate roughly half of annual revenue. Miss a week of winter storms due to maintenance and you've lost a disproportionate chunk of the year.

A word on methodology, because this matters more than people think. The EUR 80/MWh figure assumes you know all 24 hourly prices in advance and pick the perfect 4 cheapest and 4 most expensive hours — "perfect foresight." No real battery achieves this. A more realistic benchmark: a fixed-schedule battery charging at 1–5am and discharging at 5–9pm captures EUR 63/MWh gross over 2024–2025, roughly 31% less than perfect foresight. The gap between these methods is the "optimization premium" — the value of having good forecasting and flexible dispatch. A well-operated battery with modern optimization likely falls between the two, closer to the smart end. Our deep dive into spread methodology is available in the interactive spread explorer.

The time period also matters enormously. The data splits into three eras: pre-crisis (2018–mid 2021, mean EUR 49/MWh), the energy crisis (mid 2021–mid 2023, mean EUR 117/MWh), and post-crisis (mid 2023–present, mean EUR 89/MWh). The crisis — 27% of the observation period — massively lifts the all-period average. Excluding it, the mean drops to EUR 68/MWh. Any projection that uses the full-period average is implicitly assuming another crisis will occur. We use EUR 76/MWh — roughly the 365-day rolling average as of early 2024 — as our revenue model starting point, while acknowledging this sits between the post-crisis EUR 89 and the ex-crisis EUR 68.

03Oil and Gas: The Price of Everything

Since gas sets the electricity price most hours, and since gas prices in Europe are loosely tethered to oil prices through long-term contracts and LNG dynamics, the whole chain starts with Brent crude and TTF gas.

Here's the uncomfortable truth about energy forecasting: nobody is good at it. In 2020, oil briefly went negative. In 2022, European gas hit EUR 340/MWh, roughly 10x the long-term average. As of early 2026, Brent sits around $70/bbl and TTF gas around EUR 35/MWh, both roughly at historical midpoints. We use these as our baseline, because pretending we can forecast them accurately would be dishonest.

What matters for a battery is not the level of gas prices but the volatility they create. Higher gas prices generally mean higher electricity prices, which means higher absolute spreads (the gap between cheap and expensive hours widens in absolute euro terms). But the relationship isn't linear, and there's a wrinkle: if gas gets very expensive, demand destruction kicks in, and if it stays cheap, the spread narrows because gas plants can afford to run at lower utilization without the pricing stress.

Our model uses an elasticity of 0.5: a 10% change in gas price produces roughly a 5% change in the daily spread. This is calibrated from the 2019–2024 data, where we saw gas prices range from EUR 8 to EUR 340/MWh and could observe the spread's response. The confidence on the elasticity is decent (C2), but the confidence on the future gas price is not (C3), which is exactly the right level of honesty here.

The scenario range: At EUR 25/MWh gas (low scenario), spreads compress by roughly EUR 5/MWh. At EUR 55/MWh gas (high scenario), spreads widen by roughly EUR 10/MWh. The battery's annual arbitrage revenue swings by about EUR 500k–1M depending on gas — meaningful but not the dominant variable.

04The Datacenter Elephant in the Room

Ireland has an unusual problem: datacenters now consume roughly 22% of total electricity demand. For context, that's more than the entire residential heating sector. And datacenter demand has a very specific profile: it's flat. Nearly constant, 24 hours a day, 365 days a year.

This matters for batteries in a counterintuitive way. A flat 24/7 load raises the floor of demand — even at 3am, there's substantial baseload from datacenters that keeps demand (and therefore prices) higher than they'd otherwise be. This compresses the peak-to-trough ratio.

Imagine two Irelands: one with datacenters, one without. In the Ireland without datacenters, nighttime demand drops much lower, gas plants shut down, prices crash further, and the spread is wider. In the real Ireland, the datacenter baseload keeps those troughs from dropping as far. We estimate this compresses spreads by roughly EUR 2–4/MWh — a modest effect, but a structural one that gets worse as datacenter capacity grows.

EirGrid projects datacenter demand growing another 15% by 2028. Every new hyperscale campus in south Dublin is, at the margin, slightly worse for battery economics. It's a small effect per campus, but they add up.

05The Renewable Buildout

Ireland's renewable electricity share (RES-E) is around 42% as of 2025, mostly onshore wind. The government's target is 80% by 2030, which is ambitious even by Ireland's standards. The realistic trajectory probably looks more like 55% by 2028 and 65–70% by 2030, based on the actual planning pipeline.

More renewables are, on net, good for batteries. Here's why: wind and solar create more zero-price (or negative-price) hours when they're abundant, and more scarcity hours when they're not. Both effects widen the spread. German data — the best proxy we have — shows that each 10 percentage point increase in renewable penetration adds roughly EUR 12/MWh to the daily arbitrage spread.

Applied to Ireland's trajectory, that means:

This is the cavalry, from the battery's perspective. But there are two forces riding against it.

The Celtic Interconnector

A 700 MW cable to France, expected in spring 2028. In theory, it imports cheap French nuclear power during Irish peak hours, compressing the peak price. We estimated a −EUR 6/MWh impact on spreads. But here's a reality check: when the 500 MW Greenlink interconnector to Wales came online in February 2025, the detectable impact on spreads was... zero. It was confounded by gas prices being up 78% year-on-year, but the signal was genuinely lost in the noise. The "interconnectors crush spreads" story hasn't played out empirically yet.

More Batteries Eating the Same Spread

This is the real threat. Ireland currently has about 83 MW of operational BESS, with a pipeline of 10 GW (though most of it is speculative). Each additional GW of batteries competing for the same spread compresses it — because all the batteries want to buy at the same cheap hours and sell at the same expensive hours, narrowing the gap.

From GB data, each additional GW of BESS reduces per-MW revenue by roughly EUR 12k/year. For Ireland's smaller market, the effect per GW is likely larger. If 2 GW gets built by 2028 (plausible), that's roughly −EUR 11/MWh on the gross spread. If 4 GW gets built (unlikely but possible by 2030), spreads compress dramatically.

The net of all this: Renewables push spreads up; BESS fleet growth and interconnectors push them down. In our central 2028 scenario, these roughly offset, leaving the gross spread around EUR 72/MWh (vs. EUR 76 today). But the confidence interval is wide: anywhere from EUR 55 to EUR 95 depending on which force dominates. This is C3 confidence, which means "we have a defensible model but wouldn't bet the farm on any single number."

06The Actual Economics, Step by Step

Let me walk you through the financial model, one assumption at a time, so you can see exactly where the money comes in and where it leaks out.

The Upfront Bill: EUR 34 Million

Building a 50 MW / 200 MWh battery costs roughly EUR 34.0 million, or EUR 170/kWh. Here's where it goes:

A few things worth noting. The cells themselves are only 20% of the total cost — the "battery prices are falling" narrative, while true ($40/kWh FOB, down from $150 a few years ago), is less helpful than it sounds because the other 80% (grid connection, transformer, construction, permitting) doesn't follow the same learning curve. The grid connection at EUR 4.5M is highly site-specific; it could be EUR 3M or EUR 6M depending on how far you are from a 110kV substation. And the transformer has a 128–144 week lead time, which is a 2.5–3 year wait just for one component.

The range is EUR 29.6–39.9M (roughly ±15%). This is C2–C3 confidence: we have vendor quotes and Irish project comparisons, but grid connection costs in particular are a roll of the dice until you get an actual offer from EirGrid.

The Revenue: Four Streams, One Reliable

A battery in Ireland can earn money from four sources. Let me be honest about the confidence level on each.

CRM capacity payments (EUR 3.0M/yr) are the only contracted revenue. EirGrid runs T-4 capacity auctions; the 2028/29 auction cleared at EUR 149,960/MW/year, a record high driven by Ireland's tight capacity margin. A 4-hour battery gets a 40% de-rating (they don't trust it to run for 8+ hours like a gas plant), so your 50 MW earns like 20 MW. This is the safest number in the model.

Wholesale arbitrage (EUR 3.5M/yr) is the buy-low-sell-high revenue. We take the EUR 76/MWh gross spread, apply a 65% capture rate (you don't perfectly time every trade), multiply by 4 hours of duration and 0.85 cycles per day. But this number deserves a deeper look, because the chain of assumptions matters.

The capture rate (65%) comes from GB benchmarks circa 2019–2020. Here's the uncomfortable trend: our own backtest shows capture rates declining — from 72% in 2020 to 54% in 2025. The reason is structural: as more batteries and flexible assets enter the market, the easy arbitrage gets competed away and the remaining value concentrates in harder-to-predict hours. If the real capture rate is 55% instead of 65%, arbitrage revenue drops by ~15%, which is EUR 500k/year less than our projection. The spread audit flags this as the single highest-severity risk in the revenue model.

The cycling rate (0.85/day) accounts for days when spreads are too thin, maintenance downtime, and the reality that 4-hour batteries can't always execute a full cycle. But this assumption conceals an interesting upside: optimal dispatch with multiple partial cycles per day. More on this in a moment.

Where does EUR 76/MWh really come from? The 365-day rolling average of "best any 4 hours" hit EUR 76.24/MWh on January 1, 2024. By January 2025, it had risen to EUR 84.33/MWh; by January 2026, EUR 98.46/MWh. So the EUR 76 figure was already stale by the time we started using it. Our spread audit recommends the 24-month median (EUR 62/MWh) as more representative than the mean, and a 5-year ex-crisis baseline of EUR 55–60/MWh for conservative planning. We keep the EUR 76 as our central case because it roughly splits the difference between the post-crisis EUR 89 and the longer-run ex-crisis EUR 68, but you should know it sits on the optimistic side of defensible.

DASSA system services (EUR 2.0M/yr) is the number that keeps me up at night. This replaces the current DS3 programme (which pays EUR 58–300k/MW/year via regulated tariffs and expires September 2027). DASSA moves to competitive auctions starting May 2027, and no one knows what the clearing price will be because no auction has been held yet. We use EUR 40k/MW as an industry consensus estimate, but this is C4–C5 confidence — effectively a guess. It could be EUR 25k (if lots of batteries compete the price down) or EUR 60k (if EirGrid desperately needs services and few qualify).

Balancing market (EUR 1.0M/yr) is a plug number based on GB data showing that balancing represents roughly 30% of total BESS revenue. We haven't modeled this analytically. It could be significantly higher — imbalance prices in Ireland range from EUR −273 to EUR +1,453/MWh — but we'd rather underestimate than build a case on unmodeled upside.

Optimal Dispatch: The Hidden Upside

The revenue model above assumes one cycle per day: charge once, discharge once. But Irish wholesale prices don't always have just one valley and one peak. Many days have two or more tradeable windows — a morning trough-to-peak and an afternoon-to-evening trough-to-peak. What if the battery could charge, partially discharge, recharge, and fully discharge in the same day?

We modeled this properly using a state-of-charge (SoC) aware dynamic programming approach with perfect foresight. At each interval, the battery can charge at 50 MW, discharge at 50 MW, or idle, with SoC tracked from 0 to 200 MWh. This finds the true theoretical maximum — not just two discrete buy-sell trades, but any combination of partial charges and discharges.

The optimal dispatch captures ~16% more revenue than simple 4h single-cycle trading on average, rising to 20–21% in recent years as intraday price volatility increases with renewable penetration. On a typical day, the DP finds 5–8 charge and 5–8 discharge intervals — the battery is doing multiple partial cycles, buying in every cheap interval and selling in every expensive one.

This is the signature of growing intraday volatility from renewable intermittency — the price profile is getting spikier, not smoother, which creates more windows to trade. In 2019, optimal dispatch added 16% over single-cycle; by 2025, it's 20%.

After applying a realistic 60% capture rate, the uplift comes to roughly EUR 400–700k/year over baseline. It's not transformative, but it's free upside from smarter scheduling, and it gets better every year as renewables grow.

Price Impact: Can the Battery Move the Market?

A valid worry: if a 50 MW battery charges during the same cheap hours as everyone else, does its own demand push prices up enough to erode the spread? We analyzed a full year of SEMOpx bid-ask curve data (March 2025 to March 2026) to answer this.

The result is reassuring: a 50 MW battery shifts the clearing price by a median of EUR 0.80/MWh. Relative to an average spread of EUR 80/MWh, that's about 1% of the value. The Irish day-ahead market clears roughly 5–6 GW of demand per hour; a 50 MW battery is noise. Even at 100 MW, the price impact would be under EUR 2/MWh. This is good news: the arbitrage numbers in the revenue model aren't materially degraded by the battery's own market participation. (At fleet scale — multiple GW of batteries all chasing the same hours — the story is very different, which is why the BESS cannibalization analysis matters.)

The Cost Side: Where It All Goes Wrong

Here's the part that the pitch decks don't emphasize. Your battery needs to pay grid charges to the transmission system operator (EirGrid) and distribution system operator (ESB Networks) every time it imports electricity from the grid. These charges are called D-TUoS — Distribution Use of System — and they are brutal.

Look at that table for a moment. EUR 3,476k/year goes to D-TUoS charges — that's 84% of total operating costs and 41% of your gross revenue. Before you pay for maintenance, insurance, or debt service, more than a third of every euro you earn goes to the grid operator as a toll for the privilege of storing electricity.

The single largest line item is the D-TUoS System Services charge at EUR 1,971k/year (EUR 27/MWh on every MWh you import from the grid). This charge was designed to fund grid stability services — the kind of thing that batteries themselves provide. There's a reasonable argument that charging batteries this fee is like charging firefighters an arson tax. This will be important later.

Putting It Together: The 20-Year Cash Flow

Year 1 looks like this:

Year 1 looks marginally positive — EUR 1.8M free cash flow. But batteries degrade. We model 1.5%/year capacity degradation, which means your revenue drops each year while your D-TUoS capacity charges stay fixed (they're based on rated MW, not actual throughput). By Year 5, the project tips into negative annual cash flow. Over 20 years, the cumulative undiscounted cash flow is −EUR 21.4M.

The resulting metrics:

07What Could Go Wrong (and Right)

Let me be specific about the things that keep this analysis from being a simple yes-or-no answer.

Things That Could Make It Worse

BESS fleet cannibalization is the structural threat. If Ireland builds 3–4 GW of batteries by 2030 (the pipeline says 10 GW, reality says probably 2–3 GW), arbitrage spreads compress significantly. Our 2028 scenario with 2 GW of BESS shows a median spread of EUR 28.7/MWh — a 54% decline from current levels. At 2030 with more fleet, the median drops to EUR 11.9/MWh (81% decline). The battery market has a self-defeating dynamic: the more batteries that get built, the worse the economics for each one.

DASSA pricing collapse is the immediate threat. If competitive auctions drive system services payments to EUR 25k/MW (the low end of GB benchmarks), that's EUR 750k/year less than our base case. Combined with arbitrage compression, the project goes from "bad" to "catastrophic."

Grid connection surprises are the wild card. Our EUR 4.5M estimate has a range of EUR 3–6M, but in practice, EirGrid's grid connection offers have occasionally come in at multiples of initial estimates. Each extra EUR 1M of capex knocks roughly 0.3pp off the IRR.

Things That Could Make It Better

The renewable buildout exceeding expectations is the bull case. If Ireland actually hits 80% RES-E by 2030, the spread-widening effect could add EUR 30+/MWh to arbitrage opportunities, easily overwhelming the BESS compression effect. The question is whether Ireland's planning system and grid infrastructure can deliver that pace — historically, they haven't.

Higher CRM prices are possible. The T-4 2028/29 auction cleared at a record EUR 149,960/MW. If Ireland's capacity margin stays tight (which the coal exit and slow gas build suggest), future auctions could sustain or exceed this level. Each EUR 10k/MW increase in CRM clearing adds about EUR 200k/year to revenue.

Balancing market upside is real but unmodeled. GB batteries are earning 30–40% of revenue from balancing; our EUR 1M/year estimate may be conservative.

Optimal dispatch upside is modest but growing. A SoC-aware optimal dispatch adds ~16–20% to arbitrage revenue (EUR 400–700k/year), and this uplift has grown from 11% (2018) to 20% (2025) as intraday volatility increases with renewable penetration. The additional partial cycles become more valuable over time, not less. Meanwhile, a 50 MW battery's own price impact is negligible (EUR 0.80/MWh median) — you're not eroding your own spread.

The sensitivity ranking (impact on IRR):
1. D-TUoS reform: 14.3pp swing (dominant)
2. Wholesale spread (EUR 40–85): 7pp swing
3. Cycling rate (0.70–1.00): 5pp swing
4. BESS fleet growth: 5pp swing
5. DASSA pricing: 3pp swing
6. Capex (EUR 157–197/kWh): 3pp swing
7. CRM clearing price: 2pp swing

Notice something? The single largest variable — by a factor of 2x over the next biggest — is a regulatory decision, not a market outcome. Which brings us to the elephant in the room.

08The D-TUoS Problem

Let me explain why a single line item in a tariff schedule is the difference between a dead project and a viable one.

D-TUoS System Services charges were designed to fund the cost of keeping the electricity grid stable — frequency response, voltage support, inertia, and so on. These services are provided by generators, and the cost is recovered by charging all users of the distribution network a per-MWh fee on their imports.

Here's the problem: a battery storage system is both a consumer of grid electricity (when charging) and a provider of the very grid services that D-TUoS is designed to fund. Under current rules, the battery pays EUR 27/MWh on every MWh it imports — roughly EUR 1.97M/year — to fund services that it could itself provide and be paid for providing through the DS3/DASSA programme. It's paying twice: once through the charge and once through the revenue it forgoes by having to price in this cost.

This isn't just our opinion. The Electricity Storage Ireland (ESI) trade group and the European Commission's Clean Energy Package both argue that storage should be treated as a grid asset, not a consumer. The CRU (Ireland's energy regulator) has the power to grant an exemption, and there have been signals that they may do so. But no formal decision has been made.

The numbers are stark:

EUR 1.97M/year. That's the value of one regulatory decision. Over 20 years, it's the difference between losing EUR 21M and earning EUR 10M. The total swing — EUR 33M — is approximately equal to the entire capital cost of the project.

This is not a market risk. This is a policy risk. The economics of battery storage in Ireland are not determined by oil prices, gas markets, renewable buildout, or battery technology curves. They are determined by whether a regulator checks a box on a tariff schedule. Everything else is second-order.

What's the probability of reform? Honest answer: we don't know. The ESI/ECA lobbying effort is well-organized. The EU Clean Energy Package provides legal cover. The CRU has been sympathetic in public statements. But Ireland's regulatory process is slow, and "sympathetic public statements" are not the same as "published final decision." Our rough estimate is 40–50% probability within the next 5 years, which is precisely the kind of probability that makes an investment committee uncomfortable.

09The Verdict

Here are the conditions that would need to be met before committing capital:

Hard gates (all must be met):

1. CRU formally adopts D-TUoS System Services exemption for energy storage (not just "signals" — a published decision)
2. Grid connection offer at acceptable shallow cost (< EUR 5M)
3. DASSA programme achieves operational status (currently rated "red")

De-risking signals (informative, not blocking):

1. T-4 2029/30 CRM auction confirms sustained high clearing prices (auction date: 26 March 2026)
2. Post-SDP trading revenue data published for operational Irish BESS projects
3. Celtic Interconnector commissioning data showing actual spread impact

Recommended strategy: Limit pre-investment spend to < EUR 500k (site identification, planning pre-application, transformer reservation deposit). Monitor the CRU tariff review cycle. Be ready to move quickly if the exemption is granted — the combination of record CRM prices, growing renewables, and D-TUoS reform could create a brief window where battery returns are genuinely attractive before fleet growth compresses them again.

The irony of battery storage in Ireland is that the technology works, the market need exists, and the government says it wants 2+ GW of storage. The only thing stopping it from being a good investment is a line item in a tariff schedule that charges batteries for the very service they provide. Fixing this is straightforward. Whether it happens, and when, is the EUR 33 million question.

For the full assumption dependency graph showing how every number flows into every other number, see the interactive assumption map. For confidence ratings and known gaps, see the research overview. For the raw data behind every claim, see the dashboard and its linked reference pages. For interactive spread analysis by window size, year, and methodology, see the spread explorer. For historical price patterns and seasonal trends, see the price viewer.