CERES POWER HOLDINGS PLC (CWR) — Research Note

Executive summary

Ceres Power is an asset-light UK developer of solid-oxide fuel-cell (SOFC) and electrolyser (SOEC) technology that licenses its IP to major industrial OEMs (Doosan, Delta, Denso, Weichai, Thermax, Shell). Across the period covered, revenue has been volatile and lumpy — driven by upfront licence fees — with FY25 revenue down 37% YoY to £32.6m, operating losses widening to £47.6m, and the Bosch SOFC programme terminated in Feb 2025. The single most important valuation point is that royalty income, the long-term value driver of the licensing model, has only just begun (FY25 royalties: £110k from Doosan) — the current £1.2bn market cap requires that the royalty ramp materially exceeds the disclosed near-term trajectory.

Fair value estimate

• Fair value range: ~250p – 500p per share (implied market cap range £490m – £980m).
• Methodology: blended approach — (i) probability-weighted DCF on the licensing-royalty ramp using the 2030 BloombergNEF-cited SOFC market (~22GW) sized to Ceres' realistic share, (ii) cross-check against EV / forward licence-revenue multiple at 10–15x on the £45m contracted FY26 base, and (iii) sense-check against net cash floor (£83.3m at YE25).
• Key assumptions:
  • FY26 revenue ~£45m (management-disclosed contracted), gross margin 70%.
  • Operating costs reduced ~20% in 2026 per plan but still loss-making to 2028.
  • Royalty stream ramps from £110k in 2025 to a £30–60m range by 2030 (Doosan, Delta, Weichai, Denso all reaching mass production), with high (~95%) incremental margin.
  • Sustaining cash burn of £15–25m p.a. through 2026-2027 with breakeven achievable 2028–2029.
• Vs current £1,205.9m: mid-point ~£735m → downside ~ -39%, low-end downside ~-59%, high-end downside ~-19%.

Sector context

ICB Industry classification (Energy) is technically correct but misleading — this is a green-tech/industrial-technology IP licensor, not a hydrocarbon energy producer. Quality (asset-light, sector-leading 70%+ gross margin) is above typical Energy peers, but persistent losses and unproven royalty stream are below the listed clean-energy growth-stock benchmark. Closest listed peers: Bloom Energy (BE.US) — direct SOFC competitor for data-centre power, profitable scale player; ITM Power (ITM.L) — UK-listed PEM electrolyser peer; Plug Power (PLUG.US) — broader hydrogen-economy comparator (all loss-making clean-energy plays).

Investment thesis

1. AI data-centre power thesis is real and structural. Management explicitly cite Microsoft/Google/OpenAI/Blackstone UK data-centre announcements driving demand for fast-to-power, high-efficiency stationary generation. Wait times of 5–15 years for grid/gas turbines/SMRs make SOFCs a credible "bridge" technology. 2026-03 FY25 results; 2025-09 interim
2. Asset-light licensing model has extreme operating leverage if royalties scale. Doosan's 50MW factory is in production; Delta committed £170m to its Taiwan facility; Weichai signed a new manufacturing licence in Nov 2025. First royalties booked in 2025 marked the inflection. Incremental royalty income flows at near-100% margin onto a largely fixed cost base. 2026-03 FY25 results
3. Strong balance sheet for the runway. £83.3m net cash (no debt) plus 20% opex cut in 2026 provides ~3-4 years' runway against current burn rates, reducing dilution risk while the royalty stream develops. 2026-03 FY25 results

Key risks

1. Royalty ramp is unproven and consistently slower than guided. Bosch terminated SOFC activities in Feb 2025 after years of investment (€400m+ committed); China JV with Bosch/Weichai never closed; FY24/25 revenue was 37% lower YoY as upfront licence fees rolled off. The transition from licence-fee revenue to royalty-driven revenue has been repeatedly pushed back. 2025-03 FY24 results; 2024-01 trading update
2. Going-concern depends on continued new partner wins. FY25 operating loss of £47.6m is 14× FY25 revenue. Stress tests in the going-concern statement explicitly require "a slower intake of future licensee partners" mitigation. Hydrogen electrolyser FIDs have stalled industry-wide. 2026-03 FY25 results, Going Concern note
3. SOFC competition is intensifying. Bloom Energy is already at scale supplying AI hyperscalers in the US; data-centre operators may favour proven Bloom units over Ceres-licensed Doosan/Delta units which are still in factory commissioning. SOFCs also compete with gas turbines, SMRs and on-site renewables. not explicitly disclosed but inferred from competitive landscape

Operating leverage

This is a textbook high-operating-leverage business model: 70% gross margin (sector-leading), £70.1m of FY25 operating costs that are 80% fixed (R&D £48.6m, admin £14.2m, commercial £7.3m, all largely headcount-driven). Capacity utilisation of the engineering/test infrastructure is well below planned long-run levels — the FY25 plant remained pilot-scale, with the partner factories (Doosan 50MW, Delta multi-hundred-MW, Weichai TBD) carrying the volume scaling. A 10–20% revenue beat from new licence wins would drop almost entirely to EBITDA (£3-£10m incremental). However, the real leverage event is the royalty stream: if FY28 royalties hit, say, £40m vs current £0.1m base, the contribution at ~95% incremental margin would add ~£38m of operating profit — multiples of current group profit (which is deeply negative). Management's 2026 outlook of "contracted revenue £45m" plus 20% cost cut is the proximate test of operating leverage. 2026-03 FY25 results financial review

Value-trap signals

• Revenue declined 37% YoY in FY25, and recurring licence-fee revenue (the prior business model) has been falling.
• Bosch withdrawal (Feb 2025) — a flagship strategic partner that received €160m EU IPCEI funding terminated SOFC activity entirely.
• Repeated guidance misses on partner timing: China JV delayed and eventually abandoned; Bosch slipped multiple times.
• 20% AGM vote against remuneration report (2025 AGM) signals shareholder dissatisfaction.
• Prior-year accounting restatements (April 2024) over revenue-recognition timing and dilapidation provisions raised governance questions.
• Hydrogen FID delays industry-wide drag on the SOEC growth narrative.

Earnings vs expectations

The track record across the filings is more misses than beats:

• FY23: Guided "approximately double 2022 revenues" in Jan 2024 trading update; result was £22.3m vs prior £19.8m (i.e. only +12%) — a clear miss, exacerbated by the failure of the China JVs to close.
• FY24: Original guidance "broadly similar to 2024" later raised to £50-60m on Delta deal; final result £51.9m matched the lower end; then a March 2025 last-minute revenue-recognition adjustment knocked revenue from £55-60m back to £51.9m — effectively a soft miss against the upgraded number.
• H1 25 / FY25: Sep-25 guidance "around £32m" was met (£32.6m); however this was itself a downgrade from earlier implied expectations. The Bosch termination in Feb-25 had not been anticipated.
• 2026 outlook: contracted £45m disclosed — a recovery vs FY25 but still well below FY24 levels. Pattern: management consistently guides on partner-signing milestones that slip, with revenue lumpy and several large investor-disappointment events (Bosch, China JV, accounting restatement).

Conviction

3 — Moderate. Anchors: clean audited disclosure, transparent licensing-model economics, observable cash burn and runway, a clear and quantifiable royalty inflection point. Limits: the valuation is dominated by uncertain royalty trajectory (could plausibly be £10m or £100m by 2030); SOFC market share assumptions in BloombergNEF's 22GW-by-2030 forecast are themselves wide; prior-year restatement creates some hesitation on aggressive forecasting; one-off licence-fee revenue lumpiness makes near-term multiples unreliable. A different methodology (real-options/venture pricing) could justify a much higher value if the AI data-centre TAM is the dominant variable, while a peer multiple on £45m contracted FY26 revenue would suggest a much lower value.