Spread Outlook — What's Changing?

How was this page built?

Historical Baseline

Computed average 4-hour spread (charge at 01-04, discharge at 16-19) from 7 years of SEM data. Baseline: EUR 62/MWh average, EUR 76/MWh gross (before RTE losses).

Waterfall Decomposition

Built a 4-driver model estimating how future supply/demand changes affect spreads. Each driver sized independently: Offshore wind (+EUR 15-25, widens spreads by depressing off-peak), Solar (+EUR 5-10, creates midday trough), BESS fleet (–EUR 11/GW added, compresses through arbitrage), Celtic IC (–EUR 5-12, price coupling).

Greenlink Natural Experiment

Used the Feb 2025 commissioning of Greenlink (500 MW IC to Wales) as a natural experiment. Compared pre/post price correlation between SEM and GB market to calibrate the IC compression effect.

Cannibalisation Curve

Modeled the self-defeating feedback loop: more BESS → more arbitrage → spreads compress → less revenue per MW. Calibrated from Australian NEM data where BESS fleet has grown from 0 to 2+ GW.

Limitations: The waterfall drivers interact (e.g., more wind + more BESS has a different effect than either alone). We model them independently, which may overstate the net effect. The cannibalisation curve is based on Australian data — Ireland's smaller market may cannibalise faster. Timeline uncertainty on offshore wind is huge (Irish planning has historically been very slow).

Nerd level:

01 The Four Drivers Waterfall from baseline to central 2030

Spread Waterfall Baseline EUR 76/MWh to Central 2030 EUR 84/MWh

C2 — RES-E widening C3 — Celtic IC, BESS cannibalisation C3 — Gas elasticity

Each bar shows one driver's impact on the 4-hour mean daily spread between the 2024-25 baseline (EUR 76/MWh) and central 2030 projection. RES-E widening (+28 EUR/MWh from 42% to 65% RES-E) more than offsets the combined compression from Celtic IC (-6), BESS cannibalisation (-11 from +1 GW), and lower gas prices (-3). Net result: spreads widen slightly in the central case.

Driver Impact Summary

Driver	Direction	Impact (EUR/MWh)	Mechanism	Confidence
Celtic IC (700 MW)	Compress	-6	Imports cheap French power during Irish peaks; 12% of peak demand at 60% utilisation	C3
BESS cannibalisation	Compress	-11 per GW	Batteries buy low / sell high, flattening the price profile. First 2 GW: -11/MWh each	C3–C4
RES-E widening (55% to 65%)	Widen	+12 per 10pp	More zero-price hours (surplus) + more gas-set scarcity hours (droughts)	C2–C3
Gas elasticity	Varies	~0.5 elasticity	10% gas price change produces ~5% spread change; higher gas = wider spreads	C2

Net central = EUR 84/MWh. The surprise finding: spreads don't collapse in the central case. RES-E widening (+28 EUR/MWh from 42% to 65%) approximately offsets the combined compression from BESS (-11), Celtic IC (-6), and gas (-3). This holds as long as the BESS build rate stays below EirGrid's 4.7 GW target.

02 The Greenlink Experiment 500 MW interconnector, Feb 2025 — the best natural experiment we have

Before/After Greenlink Monthly mean 4hr spreads, EUR/MWh

C2 — Real SEM data; interpretation confounded by gas prices

Greenlink (500 MW to Wales) went live February 2025, increasing Ireland-GB interconnector capacity by 50%. Month-by-month comparison shows no consistent compression pattern.

71.3

Pre-Greenlink mean spread (EUR/MWh)

Feb-Dec 2024, before 500 MW interconnector

76.0

Post-Greenlink mean spread (EUR/MWh)

Feb-Dec 2025, after 500 MW interconnector

+7%

Actual change in spreads

Zero detected compression; confounded by Q1 2025 gas spike (TTF EUR 50 vs EUR 28)

Key Finding: The "15-20% Celtic IC Compression" Claim Has No Quantitative Basis

The financial model's sensitivity analysis assumes Celtic IC compresses spreads by 15-20%. This figure is not supported by evidence. C2

Greenlink precedent: 500 MW (similar mechanism to Celtic IC) showed zero detectable spread compression over 11 months of operation. The signal is lost in gas-price noise.
Back-of-envelope check: At realistic 60% utilisation during peaks, Celtic IC displaces ~7% of peak demand, producing ~7-14% peak price reduction. This supports a 5-12% spread compression range, not 15-20%.
France is getting more volatile: 436 negative-price hours in 2025, nuclear daily swing of 6 GW (up from 1.5 GW in 2022). The "flat French baseload compresses Irish volatility" narrative is outdated. C2–C3

Revised estimate: -5% to -12% spread compression (central: -8%). The price-suppressing direction is clear (France is cheaper); the volatility-suppressing direction is uncertain (France is itself becoming more volatile). Using 15-20% as the central case overstates compression. Source: SEM hourly data, SPREAD-DRIVERS.md Section 1.

03 BESS Cannibalisation The largest single risk to battery arbitrage

GB BESS Revenue Decline with Fleet Growth Revenue per MW vs installed capacity

C2 — GB empirical data (Modo Energy, LCP Delta, Rabobank)

GB BESS revenue peaked in 2021 when Dynamic Containment paid at the GBP 17/MW/h price cap. The cliff occurred when fleet capacity (~2.5-3 GW) exceeded frequency response market depth (~1.6 GW). Since 2024, revenues have partially recovered as the revenue mix shifted from ancillary services to wholesale/BM trading. Note: 2024-25 recovery makes the "71% decline" headline misleading.

Ireland Scaling: Saturates ~10x Faster Per GW

Ireland's peak demand (~6 GW) is roughly 9x smaller than GB (~55 GW). Each GW of BESS added to the Irish market has proportionally greater impact on spreads. C3–C4

BESS Fleet	Cannibalisation per GW	Basis
First 2 GW	-11 EUR/MWh	GB wholesale slope, Ireland-adjusted for smaller market
2-4 GW	-8 EUR/MWh	Diminishing returns; spread already partially compressed
4+ GW	-5 EUR/MWh	Near floor; renewable volatility creates minimum spread

~1 GW

Operational BESS fleet

~730 MW actively trading; rest on DS3 contracts

83 MW

Under construction

Tiny fraction of the 10 GW development pipeline

10 GW

Development pipeline

Largely on paper; grid connection queues and planning delays

128-144

Transformer lead time (weeks)

Natural speed limit on deployment; ~2.5-3 years from order to energisation

The bottleneck is physical, not financial. Transformer lead times of 128-144 weeks impose a natural speed limit on BESS deployment. Even with full planning approval and grid connection offers, the supply chain constrains how fast the fleet can grow. This is why the central scenario assumes only +1 GW by 2030 (to ~2 GW total), far below EirGrid's 4.7 GW target. Source: SPREAD-DRIVERS.md Section 5.1.

04 Renewable Widening The counterforce that keeps spreads alive

Negative-Price Hours in Ireland Actual and estimated trajectory

C1 — 2022-2025 actual SEM data C3 — 2030 estimated range

More renewables mean more hours where supply exceeds demand, pushing prices to zero or below. These hours are when batteries charge cheaply. The trajectory from ~20 hours (2022) toward 200-500 hours (2030 at 70% RES-E) widens the daily spread that batteries exploit.

German Comparison: What Higher RES Looks Like

Germany at ~55% RES (2025) is the closest analogue for where Ireland is heading. The results are striking. C2

125

EUR/MWh intraday spread (Germany, 55% RES)

Early 2025; up from EUR 30/MWh at 35% RES (2015-2020)

EUR/MWh intraday spread (Germany, 35% RES)

Historical baseline before high renewable penetration

Each 10pp RES-E increase adds EUR 10-15/MWh to spreads. This coefficient is derived from multiple sources: Irish academic studies (Denny et al.), observed German EPEX data, and SEM negative-price hour trajectories. Ireland's trajectory: 42% RES-E now, 70% by 2030, 80% by 2035. At +12 EUR/MWh per 10pp (central), this means +34 to +46 EUR/MWh of spread widening by 2030-2035 — the single largest positive driver for battery economics. C2–C3

Ireland RES-E Trajectory

Year	RES-E Share	Change from 2025	Spread Widening (central)	Status
2025 (now)	42%	—	—	Baseline
2030	70%	+28pp	+34 EUR/MWh	Government target
2035	80%	+38pp	+46 EUR/MWh	Government target

05 Three Scenarios (2028-2040) Central, Bear, and Bull projections for gross 4-hour spread

Gross 4-Hour Spread Projection EUR/MWh, three scenarios

C3 — Derived from C1-C2 historic data combined with C3-C4 forward assumptions

Shaded area shows the full range between bull and bear. The dashed horizontal line marks the approximate 8% WACC breakeven spread level (~EUR 55/MWh gross, requiring ~EUR 36/MWh captured). Central case remains comfortably above breakeven throughout. Bear case falls below breakeven by 2029 and continues declining. Bull case rises steadily as renewable volatility overwhelms limited storage capacity.

Scenario Summary Table

Year	Bear	Central	Bull	Range Width
2028	54.8	85.6	102.0	47.2
2029	38.6	83.1	108.8	70.2
2030	30.4	83.6	115.5	85.1
2032	27.6	84.0	120.5	92.9
2035	24.9	84.6	128.0	103.1
2038	22.0	83.4	130.7	108.7
2040	20.0	82.6	132.5	112.5

All values in EUR/MWh. Range width expands over time as scenario assumptions diverge further. Bear scenario is partly self-contradictory: at EUR 20/MWh, new BESS construction would stop, preventing further cannibalisation (self-correcting mechanism).

06 What Drives Each Scenario Key assumptions behind the three paths

Bear Scenario

10-15% probability

"BESS flood — too many batteries, not enough renewables"

BESS by 2030

5 GW

RES-E by 2030

55%

Gas Price

EUR 25/MWh

2030 Spread

EUR 30/MWh

Aggressive BESS deployment (5 GW by 2030, up from 1 GW today) crushes arbitrage spreads. Slow renewable build means limited offsetting volatility. Low gas prices further suppress the peak-price end of the spread. Revenue falls below EUR 25k/MW/yr by 2030 — well below the practical viability floor.

Self-limiting: At these revenue levels, new BESS projects become uneconomic. Construction stalls, preventing the fleet from reaching 5+ GW. The real trajectory would bend upward. This scenario is a transient stress case, not a permanent state. C3

Central Scenario

50-60% probability

"Balanced forces — renewables widen what BESS compresses"

BESS by 2030

2 GW

RES-E by 2030

65%

Gas Price

EUR 35/MWh

2030 Spread

EUR 84/MWh

BESS fleet grows modestly (+1 GW by 2030) due to transformer supply constraints and planning delays. RES-E rises to 65%, generating spread widening that offsets BESS compression and Celtic IC. Gas remains at moderate levels. Net result: spreads roughly stable at EUR 83-86/MWh through 2040.

Key dependency: This outcome requires the BESS build rate to remain far below EirGrid's 4.7 GW 2030 target. With only 83 MW currently under construction and 128-144 week transformer lead times, this is the most likely trajectory. Revenue at ~EUR 67k/MW/yr (gross arbitrage) is 15% below the financial model's EUR 80k assumption, primarily due to a lower cycling rate (0.85 vs 1.0 cycles/day). C3

Bull Scenario

20-30% probability

"German pattern — renewable volatility overwhelms storage"

BESS by 2030

1.5 GW

RES-E by 2030

72%

Gas Price

EUR 45/MWh

2030 Spread

EUR 116/MWh

BESS fleet stalls near 1.5 GW (bottlenecks persist) while RES-E surges to 72% by 2030 and 88% by 2040. Ireland follows the German pattern where rapid renewable build-out overwhelms limited storage. Celtic IC delayed to 2029 and may boost volatility by coupling to France's increasingly volatile market. Higher gas prices amplify peaks.

Early-mover advantage: Revenue exceeds EUR 100k/MW/yr from 2034 onwards. Batteries built in the next 2-3 years would earn premium returns before competition catches up. This scenario requires both stalled BESS construction AND aggressive RES-E build — plausible given current bottlenecks and government renewable targets. C3

Scenario Assumptions Comparison

Assumption	Bear	Central	Bull
BESS fleet by 2030	4-5 GW (aggressive build)	2 GW (bottleneck-constrained)	1.5 GW (stalled)
RES-E by 2030	55% (slow build)	65% (moderate)	72% (fast track)
Gas price (TTF)	EUR 25-28/MWh	EUR 35/MWh	EUR 40-45/MWh
Celtic IC	On time (spring 2028), -10 EUR/MWh	On time, -6 EUR/MWh	Delayed to 2029, -3 EUR/MWh; boosts volatility coupling
Cannibalisation regime	Aggressive (-13/GW)	Central (-11/GW then diminishing)	Minimal (fleet too small)
2030 gross spread	EUR 30/MWh	EUR 84/MWh	EUR 116/MWh

Data Sources

SEM hourly price data (spread_analysis_24month.json, Jan 2024 - Jan 2026)
Modo Energy: GB BESS Revenue Evolution, ERCOT/CAISO annual reports, France battery storage research
LCP Delta, Rabobank, Cornwall Insight: GB BESS revenue benchmarks
Timera Energy: Ireland interconnector expansion, BESS revenue rankings
Di Cosmo et al. (2019): France-Ireland interconnector modelling (Springer Energy Systems)
Bjorndal et al. (2025): NordLink price convergence (NHH Discussion Paper)
National Grid: Interconnector consumer benefits (Jan 2026)
EirGrid AIRAA 2026-2035; Wind Energy Ireland; SEAI Energy in Ireland 2024
FfE / Fraunhofer: European Day-Ahead Electricity Prices 2024; German EPEX Spot

All projections use an additive spread model (see SPREAD-DRIVERS.md Appendix A) where each driver's impact is summed from the EUR 76/MWh baseline. The model is transparent but simplified — real-world interactions between drivers are nonlinear. Confidence levels follow the C1-C5 scale where C1 is hard data and C5 is speculation. The central scenario is not a forecast; it is the most likely trajectory given current conditions and observed bottlenecks. Honest uncertainty is wide: the EUR 82-86/MWh range in 2030 spans only the central case; the full bear-to-bull range is EUR 20-132/MWh.