The state of long-duration storage in 2026

The phrase "long-duration energy storage" (LDES) usually means a storage asset with at least 8–10 hours of discharge at rated power. Lithium-iron-phosphate dominates 1–4 hour duty. Above 8 hours, the chemistry choice is genuinely open — and the financing landscape in 2026 looks very different to the slide-deck enthusiasm of 2022.

Five technologies are competing for serious project finance. Here is where each one actually stands.

Vanadium redox flow — the incumbent challenger

Vanadium flow batteries decouple power (the cell stack) from energy (the electrolyte tank), which makes them naturally scale to long durations at falling marginal cost. The chemistry has been commercial since the early 2010s; what is new is that the Invinity-Gamesa joint factory in Bathgate began shipping at 250 MWh/year capacity from Q3 2025.

Where it's being built:

• Energy Superhub Oxford 2 — 50 MW / 250 MWh (5-hour) cell + 2 MW / 24 MWh (12-hour) vanadium block. Operational since April 2026. The vanadium block is being used specifically for the Oxford City Council heat-pump trial overnight smoothing.
• CellCube Belfast — 4 MW / 40 MWh, providing reactive support and reserve to NIE Networks.
• Yadlamalka Solar Park, South Australia — 6 MW / 24 MWh, technically not GB but the project finance terms (12-year, 8.4% IRR, GIB sleeve guarantee) have become a template.

2026 capex bracket: £510–620/kWh at the 8-hour duty point. Round-trip efficiency: 72–78%. Cycle life: effectively unlimited if electrolyte is maintained.

The economic case is that vanadium degrades on calendar time, not cycle count, which makes its lifetime LCOS attractive for daily cycling. The economic problem is that vanadium pentoxide pricing remains volatile — $7.20/lb V₂O₅ at the start of 2026, against $4.80/lb three years earlier — and the electrolyte is roughly 35% of project capex.

Compressed-air energy storage (CAES) — Cheshire is the test case

The Cheshire salt caverns at Larne and Stublach have been the focus of three CAES project announcements since 2022. The one that matters in 2026 is the Hydrostor Cheshire project — 100 MW / 800 MWh advanced-adiabatic CAES, currently in the Strategic Spatial Energy Plan submission process, with FID targeted for Q4 2026.

A-CAES stores heat from compression separately (recovering it during expansion, hence no fuel burn), which puts round-trip efficiency at 55–62% — well below electrochemical, but the capex per kWh is dramatically lower at long durations. Hydrostor publishes £/kWh installed costs around £180–240 at 12-hour duty, falling toward £140 at 24-hour.

The bottleneck is geology. CAES requires either solution-mined salt caverns or porous-rock reservoirs, both of which constrain siting. Outside Cheshire, the next viable UK sites are in East Yorkshire and around Aldbrough; Scotland has limited suitable geology.

Iron-air — Form Energy's UK pipeline

Form Energy's iron-air chemistry stores energy by reversibly rusting iron pellets. Charge: reduce rust to iron, releasing oxygen. Discharge: oxidise iron back to rust. The reaction is slow — which is precisely why it is suitable for 100-hour duty.

Form announced a UK-Ireland pipeline in November 2025: three projects totalling 15 MW / 1,500 MWh (100-hour), all targeted at island and end-of-line resilience applications where the alternative is a diesel back-up generator. The first — on the Isle of Eigg — broke ground in March 2026.

2026 capex bracket (Form's own published number): $20/kWh at the cell level, equating to roughly £80–110/kWh installed at 100-hour duty in current UK projects. The catch is that Form's product is not for daily cycling — round-trip efficiency is around 40–45%, and the asset's economics depend on rarely (5–15 events/year) being called.

Gravity storage — quietly receding

Energy Vault's tower-and-block design generated headlines in 2022 and has generated almost no new project announcements since. The Chinese commercial tower in Rudong (25 MW / 100 MWh) has been operating since late 2023 with reported availability above 92%, but no comparable UK or EU project has reached FID, and the cost gap to vanadium and CAES has widened, not narrowed.

The cleaner gravity story is Gravitricity, which uses suspended weights in disused mine shafts. Their 250 kW Edinburgh demonstrator validated the control system in 2021; the proposed 4 MW / 8 MWh commercial unit in a Czech mine reached planning consent in 2025 but is not yet financed. Promising; not yet deployable.

Thermal storage — the wildcard

Two interesting plays:

• Antora Energy (antoraenergy.com) — carbon-block thermal storage, charged electrically and discharged as 1,800°C process heat for industry. Genuinely interesting for cement and chemicals; not really a grid asset.
• EnergyNest — thermal concrete modules at modest temperature, sold into industrial decarbonisation. Their first GB project (Sunbury, Surrey) is feeding a glasshouse complex and avoiding ~12 GWh/year of gas burn.

Both compete in the industrial-heat decarbonisation market more than the electricity market. They are LDES adjacent, not LDES core.

The financing reality

In 2026, GB long-duration storage projects above 8 hours fall into three financing buckets:

Bucket	Capex (£/kWh at 8h)	Status
Vanadium flow	£510–620	3 projects financed, 4 more in LDES Cap & Floor Round 1
CAES (A-CAES)	£180–240	1 project (Hydrostor) at advanced FID
Iron-air	£80–110 (at 100h)	3 projects, all islanded

The single biggest catalyst is Ofgem's LDES Cap & Floor scheme, modelled on the existing interconnector regime. Round 1 closed for applications in February 2026 with a £1.0 bn-£1.5 bn revenue envelope; results are expected in autumn. Without it, almost none of the above projects would be financeable in their current form.

If you want the one-paragraph summary: vanadium flow is the credible 8–12 hour incumbent at high capex; A-CAES is the cheap-per-kWh option if you have the geology; iron-air is genuinely transformative for islanded resilience applications but does not compete for daily cycling; gravity is interesting but not financed; thermal is real but mostly an industrial-heat story.

The interesting trend, if you stand back, is that the question "which LDES chemistry wins?" has stopped being interesting. The right question for 2026 is which chemistry wins at which duty point, and the answer is starting to harden:

• 4–8 hours: lithium iron phosphate, increasingly with second-life cells
• 8–16 hours: vanadium flow
• 16–24 hours: A-CAES where geology permits, otherwise vanadium
• 100+ hours, low cycle count: iron-air

The other technologies will continue to exist. They will not, on current cost trajectories, take material market share.