Batteries vs PSH: Economics and costs

From FletchWiki

Sub-page of How do batteries and pumped storage hydro compare as electricity storage technologies for renewable-intensive systems like the UK's?

Summary verdict

Battery energy storage systems (BESS) are now cheaper than pumped storage hydro (PSH) at short discharge durations (1–4 hours), and this cost advantage is widening rapidly as lithium-ion pack prices continue their structural decline. PSH remains competitive — and in most cases cheaper on a levelised basis — for discharge durations above approximately 8–10 hours, with the 4–8 hour range genuinely contested. For the UK specifically, the economics are further complicated by asymmetric policy support: BESS has benefited from high merchant revenues from frequency services (volatile but historically high), while PSH projects have struggled to obtain revenue certainty for the long-duration services they are best suited to provide. A new cap-and-floor mechanism for long-duration energy storage (LDES), announced in 2024–25, is intended to correct this — but in its first allocation round it explicitly excludes batteries.

Key findings:

  • BESS capital cost has fallen ~90% since 2010; PSH capital cost has not fallen materially
  • At a 4-hour duration, BESS LCOS (~$65–$254/MWh depending on method and year) is broadly competitive with or cheaper than PSH ($140–$186/MWh on comparable bases)
  • PSH remains cheaper per kWh of energy capacity at long durations because the reservoir cost scales slowly with energy while battery cost scales linearly
  • UK BESS revenues were £37,000–£88,000/MW/year in 2024, but have been highly volatile
  • The decisive policy question is whether the UK's LDES mechanism will create long-term revenue certainty for PSH at the scale needed to justify new development

1. Capital costs

Battery energy storage systems (BESS)

Lithium-ion battery pack prices have undergone one of the fastest cost reductions of any energy technology:

Year Average Li-ion pack price ($/kWh) Source
2010 ~$1,200 BloombergNEF
2020 ~$137 BloombergNEF BNEF BATMON
2023 ~$139 BloombergNEF
2024 ~$115 (pack); ~$83 (cell) BloombergNEF 2024
2025 (est.) ~$117 (system/turnkey, global) BNEF / Lazard LCOS v10

Note: pack prices and turnkey system prices differ significantly. A fully installed UK grid-scale BESS system (including inverters, BMS, grid connection, civil works) costs approximately £250–£350/kWh (roughly £305/kWh as a central estimate for 2025 UK deployment), versus the headline pack price. The ratio of system-to-pack cost has been roughly 2–2.5x historically, but is compressing as BOS (balance of system) costs fall more slowly.

For power capacity, UK BESS capital costs are approximately £150,000–£200,000/MW for a 2-hour system, rising to £400,000–£500,000/MW for a 4-hour system.

BESS costs fell roughly 40% in 2024 alone — an exceptional single-year decline driven by Chinese LFP manufacturing scale and oversupply. LFP (lithium iron phosphate) chemistry has become dominant for grid-scale applications, displacing NMC (nickel manganese cobalt), partly for cost reasons and partly for its superior cycle life and thermal stability.

Pumped storage hydro (PSH)

PSH capital costs are highly site-specific and have not fallen materially in real terms over the past two decades. The dominant cost components are civil engineering (tunnelling, dam construction) and electromechanical equipment (turbines, generators, penstock).

Source PSH capital cost range Basis
NREL ATB 2024 $1,730–$4,500/kW Installed, US projects
IEA (2019) $500–$4,600/kW Global range
IRENA (2020) $1,050–$3,800/kW Global range
UK projects (estimated) £2,000–£5,000+/kW Coire Glas ~£1,500m for 1,500 MW = ~£1,000/kW power; but energy cost higher
NREL ATB 2024 (mid) ~$2,700/kW US reference

The power/energy cost structure of PSH differs fundamentally from batteries:

  • Power capacity cost (£/kW): driven by turbine/generator size — roughly £600–£1,500/kW
  • Energy capacity cost (£/kWh): driven by reservoir volume — typically £5–£30/kWh for durations above 8 hours, far cheaper per kWh than batteries at those durations

This means PSH has a structural cost advantage for long-duration storage on an energy-capacity basis, but is capital-intensive on a power-capacity basis. A 10-hour PSH project at £10/kWh energy cost is far cheaper per kWh than a 10-hour battery system at £250–£350/kWh.

The UK's consented Coire Glas project (SSE Renewables, 1,500 MW / 30 GWh, Upper Loch Ness) illustrates this: at ~£1.5bn estimated capital cost, the power-capacity cost is ~£1,000/kW and the energy-capacity cost is ~£50/kWh — competitive with batteries at that duration.

2. Operation and maintenance costs

Technology Fixed O&M Variable O&M Notes
Li-ion BESS ~2–2.5% of capex/year (~£5,000–£8,000/kW/year for 4h system) Minimal Includes augmentation (capacity replacement) costs; augmentation is a significant hidden cost for long-duration cycling
PSH ~$18/kW/year (NREL ATB 2024) ~$0.51/MWh Very low variable cost; long-lived civil infrastructure reduces per-year cost over 50–100 year life

A critical BESS O&M issue is battery augmentation: as cells degrade (typically 1–3%/year for LFP), some cells must be replaced to maintain rated capacity. This is not always captured in headline LCOS figures. Over a 15-year project life, augmentation can add 15–30% to total lifecycle cost.

PSH has negligible equivalent degradation cost — the reservoir and civil infrastructure are effectively permanent; electromechanical equipment has a 40–50 year life with major refurbishment at 20–25 years.

3. Levelised Cost of Storage (LCOS)

Methodology

LCOS is the dominant cost comparison metric for storage, analogous to LCOE for generation. It expresses the total lifetime cost per MWh of energy discharged:

LCOS = (Total lifecycle costs) / (Total energy discharged over lifetime)

Total costs include: capital, financing (cost of capital), O&M, augmentation, decommissioning, minus any residual value. Total energy discharged depends on: round-trip efficiency, cycle frequency, degradation, design life.

LCOS is duration-dependent: a storage asset used for 1-hour cycling has a different LCOS than the same asset used for 4-hour cycling, because the same capital cost is spread over more or less total energy depending on how often and how deeply it is cycled.

This makes direct BESS vs PSH LCOS comparison valid only at a specified duration and cycling assumption.

Current estimates

Source Technology Duration LCOS ($/MWh) Year
Lazard LCOS v10 Li-ion (utility) 4h $115–$254 2024
Lazard LCOS v10 Li-ion (utility) 1h $135–$322 2024
Lazard LCOS v9 Li-ion (utility) 4h $120–$310 2023
Ember (2025) Li-ion (utility) 4h ~$65 2025 est.
Lazard LCOS v10 PSH 8–12h $140–$186 2024
NREL (2023) Li-ion 4h ~$200–$230 2023
Schmidt et al. (2019) PSH long-duration ~$180–$220 2019 (projected to 2025)
DOE / INL Li-ion 4h 4h ~$150–$250 2024

Key observations:

  • The Ember $65/MWh figure is an outlier on the low end and likely reflects optimistic assumptions on capital cost, cycling frequency, and financing. It may represent best-case forward-looking estimates for 2025–2026 Chinese-supply projects.
  • Lazard v10 ($115–$254/MWh for 4h BESS) is the most widely cited mainstream benchmark; the range reflects differences in financing cost, technology vintage, and market.
  • PSH at $140–$186/MWh (Lazard v10) is for 8–12 hour systems. At shorter durations PSH LCOS would be higher because the capital cost is spread over fewer MWh.
  • On a like-for-like 4-hour comparison, BESS is competitive with or cheaper than PSH in most scenarios.
  • On an 8–12 hour comparison, PSH is competitive or cheaper, particularly for projects with low civil cost.

Duration crossover

The key analytical question is: at what discharge duration does PSH become cost-competitive with batteries?

Current consensus from the literature:

  • Below 4 hours: BESS almost always cheaper on LCOS basis
  • 4–8 hours: contested; outcome depends heavily on BESS cost assumptions and PSH site quality
  • Above 8 hours: PSH generally cheaper, especially as battery costs for longer-duration systems scale linearly while PSH reservoir costs do not
  • Above 12 hours: PSH has a strong structural advantage

This crossover is shifting towards longer durations over time as BESS costs continue to fall. Some analyses (e.g. BNEF 2023, Aurora Energy Research) project that by 2030, BESS may be competitive up to 6–8 hours.

4. Cost trajectories

Battery costs

The structural drivers of BESS cost reduction are:

  • Manufacturing scale (predominantly China/LFP): cell production has grown ~10x since 2015
  • Chemistry improvements: LFP has displaced NMC for most grid applications; sodium-ion and solid-state batteries are on the horizon
  • Supply chain maturation: cathode, anode, electrolyte, separator costs all declining

Projections (selected):

Source Metric 2025 2030 2035 2050
BNEF 2024 Li-ion pack ($/kWh) ~$115 ~$80–90 ~$60–75 ~$40–60
NREL ATB 2024 Utility BESS system ($/kWh) ~$280 ~$195 ~$155 ~$100
Cole & Frazier (NREL, 2023) 4h utility BESS ($/kWh) ~$270 ~$190 — ~$120
Schmidt et al. (2019, Joule) Li-ion (projected) ~$120–150 ~$80–100 — ~$50

The 2024 price drop (40% in one year) exceeded most projections. Forward projections from pre-2024 studies are likely to prove too conservative.

PSH costs

PSH capital costs are not projected to fall materially. Civil engineering and electromechanical costs track general construction inflation. There is no analogue to the manufacturing scale effect that drives battery cost reductions.

NREL ATB 2024 projects PSH costs remaining broadly flat in real terms through 2050, with modest reductions (5–10%) from improved construction techniques and project design.

The relative economics of PSH vs batteries thus shift continuously in favour of batteries over time — the only question is how fast and at what duration threshold.

5. UK-specific economics

Revenue landscape for BESS

UK grid-scale BESS operators stack multiple revenue streams:

Revenue stream Mechanism Typical value (2024) Battery suitability
Dynamic Containment (DC) Frequency response (FFR replacement) £5–£20/MW/hour Excellent (sub-second response)
Dynamic Moderation (DM) Slower frequency response £2–£10/MW/hour Excellent
Balancing Mechanism (BM) Real-time dispatch by NESO Variable; can be very high Good
Capacity Market (CM) Annual capacity auction (T-1/T-4) ~£50,000–£65,000/MW/year (2024 T-1 clearing) Yes, qualifies
Wholesale arbitrage Buy low, sell high Highly variable Good for 2–4h systems
Total (2024 estimate) — £37,000–£88,000/MW/year —

The £37,000–£88,000/MW/year range from Modo Energy and Cornwall Insight reflects extreme volatility in 2024: Dynamic Containment prices collapsed mid-year as BESS capacity grew faster than demand for the service, then partially recovered. This revenue volatility is a significant risk for project finance.

A 2-hour, 100 MW BESS project at £150m capital cost with £60,000/MW/year revenue generates £6m/year. At a 10% discount rate, the project is borderline viable — which explains why developers are highly sensitive to service price movements.

Revenue landscape for PSH

PSH is poorly served by the current UK market structure:

  • Capacity Market: PSH qualifies and typically clears at the same price as other capacity, but CM revenues alone (~£50,000–£65,000/MW/year) are insufficient to justify new PSH development (which requires £1,500–£3,000+/MW/year in equivalent annualised capex to be economic)
  • Frequency services: PSH is less well-suited than batteries for DC/DM (response time disadvantage), though it can provide slower frequency services
  • Balancing Mechanism: PSH participates but at lower frequency than batteries
  • Wholesale arbitrage: PSH's natural use case — but UK wholesale price spreads have been insufficient to make new PSH viable on merchant basis alone

The result is that no new PSH has been built in the UK since Dinorwig opened in 1984 — a 40-year gap. The consented projects (Coire Glas, Red John) have not progressed to Final Investment Decision.

The LDES cap-and-floor mechanism

In response to the PSH investment gap, DESNZ (Department for Energy Security and Net Zero) has developed a Long Duration Energy Storage (LDES) cap-and-floor mechanism, modelled on the regime used for interconnectors:

  • The mechanism guarantees a floor revenue (protecting against low market prices) and a cap (clawing back revenue above a threshold), in exchange for a commitment to build and operate
  • It provides the long-term revenue certainty that PSH requires to support project finance
  • Round 1 (2024–25) explicitly excludes battery storage — it is targeted at technologies like PSH and compressed air that cannot compete for short-duration services

This represents a deliberate policy asymmetry: batteries are supported through market revenues and the capacity market; PSH is supported through a bespoke long-duration mechanism. Whether the mechanism is well-calibrated to unlock the pipeline of consented UK PSH projects (Coire Glas ~£1.5bn, Red John ~£400m, etc.) remains to be seen.

6. The "batteries in series" question

The research strategy raised the question of whether short-duration batteries deployed at scale could simulate longer-duration storage by being used "in series" to smooth multi-hour supply gaps.

The economics are unfavourable for durations above 6–8 hours:

  • A 4-hour BESS costs ~£305/kWh. To store 8 hours of energy requires two 4-hour systems in series — costing ~£610/kWh equivalent. PSH at 8+ hours costs ~£5–£30/kWh for the incremental energy capacity.
  • In a renewable system, the need is often for energy volume (MWh) at moderate power (MW) over multi-day periods. Batteries scale linearly in cost with energy; PSH scales sub-linearly.
  • The optionality argument (same batteries can serve peak power needs or extended duration needs depending on dispatch) has some merit, but the fundamental energy-volume economics favour PSH at long durations.

The more relevant question may be: can a combination of 4-hour batteries (for short-duration, high-value services) plus PSH (for long-duration buffering) beat either technology alone? Most system modelling (e.g. Dowling et al., 2020) finds that optimal storage portfolios include multiple duration brackets rather than a single technology.

7. Summary comparison table

Parameter Li-ion BESS (4h) PSH (8–12h) Advantage at parity
Capital cost (power) £150,000–£200,000/kW £600–£1,500/kW PSH (at long durations)
Capital cost (energy) £250–£350/kWh £5–£50/kWh (>8h) PSH at long duration; BESS at short
LCOS (2024–25) $115–$254/MWh $140–$186/MWh Similar; BESS slightly cheaper at 4h
Fixed O&M ~2–2.5% capex/yr ~£18/kW/yr PSH (lower lifecycle O&M)
Design life 15 years 50–100 years PSH
Cost trajectory Falling rapidly (~-40% in 2024) Flat in real terms BESS (improving)
UK revenue certainty High (multiple services) but volatile Low; cap-and-floor emerging BESS (currently)
Crossover duration — — BESS <4h; PSH >8h; 4–8h contested

Sources

Capital and LCOS

Cost trajectories

UK market revenues

  • Modo Energy, UK Battery Storage Revenue Tracker (monthly, 2024): https://modoenergy.com/research
  • Cornwall Insight, Battery Storage Report (2024) — £37,000–£88,000/MW/year revenue range
  • NESO, Balancing Services Roadmap — Dynamic Containment / Dynamic Moderation pricing
  • Ofgem, Capacity Market Results (T-1 2024, T-4 2024) — clearing prices

UK policy (LDES mechanism)

System modelling

  • Dowling, J.A. et al., "Role of long-duration energy storage in variable renewable electricity systems," Joule 4(9), 1907–1928 (2020): https://doi.org/10.1016/j.joule.2020.07.007
  • Sepulveda, N.A. et al., "The role of firm low-carbon electricity resources in deep decarbonization of power generation," Joule 2(11), 2403–2420 (2018): https://doi.org/10.1016/j.joule.2018.08.006
  • Strbac, G. et al., Value of Flexibility in a Decarbonised Grid and System Externalities of Low-Carbon Generation Technologies, Imperial College / BEIS (2016, updated 2019)

UK PSH projects

Retrieved from "https://fletch.metakarma.org/index.php?title=Batteries_vs_PSH:_Economics_and_costs&oldid=159"