CECadence
GB battery-storage revenue forecasting — priced at the cannibalisation equilibrium
A Compounding Energy product
cadence.compoundingenergy.com
Your spreadsheet says the battery earns £X.
Once the whole GB fleet is competing, it earns £Y.
CECadence prices the difference — before you finance it.
100–300%+
naive over-statement
at 2030 fleet sizes

The problem

A battery earns from price spreads — charge cheap, discharge dear — plus frequency response, Capacity Market and Balancing Mechanism. But the GB fleet is building out fast: our own index measured ~5.8 GW of actively-traded capacity in August 2026, and our scenario set carries 12–25 GW by 2030 (18 GW central). Every new battery charges and discharges in the same hours, flattening the very spreads they all live on. That is cannibalisation: forecasts that ignore it overstate 2030 revenue badly, and a project financed on the naive number quietly breaches its covenants.

What CECadence does

It solves the whole system in equilibrium — the future fleet, how every battery dispatches, and the resulting prices — all at once, until they are mutually consistent. The headline output is the bias: how far a naive, price-taker forecast overstates the realistic, cannibalisation-aware number. That single figure is what protects a financing decision.

What you get

Cannibalisation bias

The naive-vs-equilibrium gap, per scenario and year — the number a lender needs to see.

Revenue stack

£/MW/yr split across wholesale arbitrage, frequency response, Capacity Market and BM.

Bankable finance

NPV / IRR / DSCR against the 1.40× (P50) and 1.10× (P90) covenants, with P90/P50/P10 bands.

Site & portfolio

Appraise a candidate by location and duration; compare sites head-to-head; NPV tornado drivers.

Built for

Operators
What will my fleet realistically earn over the next decade?
Developers
Which site, which duration, what NPV — and how do my options compare?
Investors & lenders
Does this revenue survive cannibalisation and still cover the debt?

The calibrated second opinion — and why you can trust it

An investment committee usually sees one forward curve, from one provider, with nothing independent to test it against. Run ours beside it and the disagreement becomes the diligence. And ours is graded, the misses published with the hits.

To our knowledge the only GB BESS forecast with a graded, pre-registered validation record — commitments in dated commits before outcomes were known.

Ask any forecast vendor for their graded record. The price is published too: a single-asset second opinion is a £12,000 one-off, creditable against a subscription taken within 90 days.

  • Ten graded backtest cells against realised GB prices 2022–2026H1, each published with its pass/fail — two are pre-registered held-out windows whose rules were frozen in dated commits before any of their data was fetched, and published unmodified
  • Held-out 2025 mean error 1.6% (realised basis); the 2026H1 realised mean (7.8%) missed our own 5% gate and is published straight — full error table in-product
  • Calibrated to our own GB BESS benchmark, built from public Elexon & NESO data — no third-party licence
  • Every run reproducible from its own manifest and audit-keyed; methodology fully disclosed in-product; the out-of-sample scarcity-tail weakness published, not hidden