Back to catalogue
№ 153 25 filings · 2021-09-30 → 2026-09-23

CERES POWER HOLDINGS PLC

CWR
Energy Share price 309p Market cap £660m Overall fit 470 /1000

Genuine AI-receiver optionality via data-centre SOFC and best-in-class operating leverage from a licensing model, but valuation already discounts a lot of the good news and the company is structurally loss-making with a history of partner slippage — partial fit for the strategy.

Fair value range 310p–545p Mid case · £910m
Absolute upside +37.9% vs current market cap
Conviction 2/5 confidence in fair call
Supports the call
  • Clear licensing economics with verifiable partner list
  • Fortress net-cash balance sheet post June 2026 raise
  • First royalties recognised — model transition is underway
Limits the call
  • Royalty rates and partner scale-up timing not disclosed
  • History of missed guidance and abandoned JVs (Bosch/China)
Methodology

Forward EV/Sales on 2028 revenue with DCF sanity check

In one line · bull case

High-operating-leverage SOFC IP licensor with credible AI-data-centre partner pipeline and £172m net cash, but valuation already prices a partner scale-up that has historically slipped.

In one line · biggest risk

Partner deployment timelines slip another 12–24 months, delaying the royalty inflection and forcing further cash burn against a valuation that already assumes it.

Drivers
AI beneficiary 55 /100
Real data-centre power thesis via Centrica/Delta partnerships but AI-driven revenue not yet material.
Operating leverage 82 /100
Asset-light licensing model with 70%+ gross margin and largely fixed cost base — classic high-leverage setup.
Earnings vs expectations 30 /100
Multiple guidance cuts and delayed JVs (Bosch, China) over the past three years.
Growth momentum 55 /100
2026 contracted revenue £45m vs 2025 £32.6m returns to growth after a down year.
Moat 55 /100
Proprietary steel-supported SOFC IP and long partner relationships, but faces Bloom Energy and other SOFC competitors.
Earnings quality 45 /100
Revenue lumpy from licence timing, prior-year restatements in 2024, and non-cash exceptionals reduce clarity.
Management quality 50 /100
Long-serving CEO delivered partner network but Bosch exit and abandoned China JV are execution blemishes.
Cyclicality 45 /100
Energy-transition capex exposure, not classical cyclical but sensitive to hydrogen/data-centre capex cycles.
Leverage 5 /100
Net cash of £172m and no material debt — a fortress balance sheet.
Value-trap signals · 4
  • Repeated JV/licence slippage (China JV never completed, Bosch withdrew after IPCEI grant)
  • Prior-period restatements disclosed in 2024
  • Customer concentration (top three = 93% of H1 2026 revenue)
  • Royalties still de minimis (<£200k) relative to embedded market expectations

CERES POWER HOLDINGS PLC (CWR) — Investment Research Note

Executive summary

Ceres Power is a UK-listed developer of solid oxide fuel cell (SOFC) and electrolyser (SOEC) technology deployed via an asset-light licensing model with partners including Doosan, Delta, Denso, Weichai, Shell, Thermax and — most recently — Centrica for UK/European data-centre power. The operating trajectory over five years has been volatile and back-loaded: revenue oscillated between £22m and £52m as one-off licence fees were recognised, adjusted EBITDA losses remained £20–33m, and cumulative accumulated losses reached £345m by June 2026, though a June 2026 £102.6m equity raise restored net cash to £172m. The single most important point for valuation today is that the market is capitalising a not-yet-material royalty stream from an AI-data-centre SOFC thesis whose economic upside depends entirely on partner execution over the next 2–4 years.

Fair value estimate

  • Methodology: revenue-scaled forward EV/Sales combined with sanity-check DCF. This is the appropriate framework because Ceres is pre-profit, licence revenue is lumpy, and royalties (the terminal driver) are not yet material.
  • Key assumptions: 2026 contracted revenue £45m 2026-09 interim; three-year revenue CAGR of ~25% (partners scaling: Doosan 50MW → 200MW target, Delta £240m factory investment, Weichai 200MW target by 2027) yielding ~£85m 2028 revenue; gross margin holding at ~70% (licensing mix); operating cost base stabilising near £60m post-transformation (14% cut from 2025 base) 2026-09 interim; net cash £172m 2026-09 interim. Apply EV/Sales of 6–12x on 2028 revenue (peers such as Bloom Energy trade at high mid-single-digit EV/S despite scale; a UK IP licensor with royalty optionality warrants a moderate premium if partner scale delivers).
  • Range: enterprise value £510m–£1,020m; adding £150m of expected 2028 net cash (allowing further burn) → market cap £660m–£1,170m → fair value ~310p–545p per share, midpoint ~425p / c.£910m market cap.
  • Vs. current £876.7m market cap at 402p: roughly fair. Absolute upside/downside on midpoint: +6%; the range spans −23% to +36%.

Sector context

The ICB "Energy" classification is technically correct but misleading — this is a clean-tech/industrial-IP licensor, not a traditional oil & gas name. Quality profile: above sector on gross margin (71% vs. sector low-single digits), above on balance sheet (net cash vs. levered energy peers), but below on cash generation (structurally loss-making). Listed peers: Bloom Energy (BE) — closest pure-play SOFC comparator, at scale and profitable at gross-margin level; AFC Energy (AFC.L) — smaller UK fuel cell peer; ITM Power (ITM.L) — UK electrolyser peer with a similarly troubled cash-burn history.

Investment thesis (3 bullets)

  • Data-centre time-to-power tailwind with quantifiable partner activity. SOFCs offer 24/7 behind-the-meter power in months, vs. 5–15 years for grid upgrades and 5+ years for gas turbines. Doosan's £60m contract with Reverion for European SOFC stacks, Delta's £240m Taiwan factory decision, and the March 2026 Centrica partnership specifically targeting UK/European data-centre and industrial power represent tangible commercial pull, not marketing 2026-09 interim.
  • Extreme operating leverage baked into an asset-light licensing model. Gross margin was 71% in H1 2026 despite lower-margin mix; the fixed cost base was cut 14% via 2025 restructuring; royalties (first recognised in 2025 at £110k, £181k in H1 2026) are near-100% incremental margin. Every incremental £10m of royalty income should drop almost entirely to operating profit 2025-12 annual; 2026-09 interim.
  • Fortress balance sheet buys time for the royalty inflection. Net cash of £172m at June 2026 following the oversubscribed £102.6m raise gives 5+ years of runway at current adjusted EBITDA loss, removing dilution/going-concern risk over the critical partner-scale-up window 2026-09 interim.

Key risks (3 bullets)

  • Partner execution risk is the entire thesis. Bosch withdrew from SOFC development in February 2025 despite prior IPCEI funding of ~€160m; the planned China three-way JV with Bosch/Weichai never completed; and the Weichai licence had to be re-signed on a new basis in Nov 2025. The record of partner slippage is real 2025-03 final results 2024; 2026-03 final results 2025.
  • Valuation requires partner scale within 2–3 years. Even the base case implies ~85% revenue growth over three years; if royalty ramp is delayed 12–18 months (as SOFC deployment lags), the shares are worth less than current price. Note H1 2026 revenue was concentrated 59%/18%/16% in just three customers — high single-partner risk 2026-09 interim.
  • Structurally loss-making with continuing cash burn. Operating cash outflow was £10.7m in H1 2026 alone; £47.5m loss in 2025; accumulated losses of £345m. Balance sheet is a runway asset, not a value driver. A second raise cannot be ruled out if royalties disappoint 2026-09 interim; 2026-03 final results 2025.

Operating leverage

This is one of the clearest cases of high operating leverage in the UK small/mid-cap universe. The cost base is overwhelmingly fixed: R&D was £18.3m in H1 2026 (down from £25.6m via transformation), administrative expenses £8.1m, commercial £4.3m — total £30.7m operating cost effectively a fixed platform servicing partners. Gross margin was 71% in H1 2026 and 70% in FY2025, and would be 80%+ on pure licence/royalty periods (79% in H1 2025). Because Ceres does not manufacture at scale — partners do — incremental revenue carries minimal COGS. A plausible +10–20% surprise on 2028 revenue (£85m base → £95–100m) at 75% gross margin adds £7–11m of gross profit against a broadly flat cost base, potentially the difference between a small operating loss and a small profit. The true inflection is royalty scale: Weichai targets 200MW by 2027, Delta targets pilot production 2028; even a low single-digit royalty rate on multi-hundred-MW deployment would flow near-100% to EBIT. The observable inflection point is the transition from licence-fee volatility to recurring royalty compounding — likely 2027–2028 2026-09 interim; 2026-03 final results 2025.

Value-trap signals

  • Repeated pattern of missed/delayed strategic milestones: China JV with Bosch/Weichai flagged as imminent for two consecutive years then abandoned; Bosch withdrawal February 2025 after €160m IPCEI funding 2025-03 final results 2024.
  • Guidance cuts: 2025 revenue guidance was originally "broadly similar to 2024" (£52m), then revised down to £32m at H1 2025 2025-09 interim.
  • Prior-year restatements disclosed in 2024 relating to revenue recognition, dilapidation provisions and capitalised costs — reduces reported earnings quality 2024-04 final results 2023.
  • Customer concentration (top three partners = 93% of H1 2026 revenue) 2026-09 interim.
  • First royalties are tiny (£110k FY25, £181k H1 26) versus the market's implicit 2028+ expectation.

Earnings vs. expectations

The record is skewed negative. FY2025: management guided "broadly similar" to £52m in early 2025 then cut to c.£32m mid-year; delivered £32.6m — miss vs. original. FY2024: originally guided £50–60m revenue at H1, refined to £55–60m in January 2025, then reset back to £50–60m in March 2025 due to a £5m milestone timing issue; delivered £51.9m — met the wider range but with confusion. FY2023: guided "around double" 2022 in trading updates; delivered £22.3m vs. £22m in 2022 — miss (offset by later Delta signing). H1 2026 came in "in line with expectations." Pattern: more misses than beats, largely driven by licence-signing/JV slippage outside the company's full control multiple trading updates 2023–2025.

Conviction

Conviction: 2 (low). The filings are clean and disclosure is good, but the fair value is a wide band with weak conviction at any single point.

  • Anchors: licensing economics are well-disclosed; balance sheet is unambiguous; partner list is verifiable and recently expanded (Centrica March 2026).
  • Limits: royalty rates are not disclosed, partner scale-up timelines have historically slipped by 12–24 months, and the difference between a £300m and a £1.5bn business two years out depends on execution I cannot verify from filings alone. This is a genuinely optionality-driven story.
Filings consulted · 29

Every document the LLM read for this note. Click any row to open the source.

  1. 2026-09-23Interim Results2026-09-23_9785577_interim-results.md0.90
  2. 2026-09-03Notice OF Interim Results2026-09-03_9752895_notice-of-interim-results.md0.90
  3. 2026-05-14Result OF Agm2026-05-14_9569356_result-of-agm.md0.30
  4. 2026-04-14Notice OF Agm Amp Publication OF Annual Report2026-04-14_9519404_notice-of-agm-amp-publication-of-annual-report.md0.95
  5. 2026-03-26Final Results For The Year Ended 31 December 20252026-03-26_9492618_final-results-for-the-year-ended-31-december-2025.md1.00
  6. 2025-09-26Interim Results2025-09-26_9133031_interim-results.md0.77
  7. 2025-09-02Notice OF Interim Results2025-09-02_9083802_notice-of-interim-results.md0.58
  8. 2025-05-15Result OF Agm2025-05-15_8880993_result-of-agm.md0.20
  9. 2025-04-11Notice OF Agm Amp Publication OF Annual Report2025-04-11_8827026_notice-of-agm-amp-publication-of-annual-report.md0.62
  10. 2025-03-21Final Results For The Year Ended 31 December 20242025-03-21_8790262_final-results-for-the-year-ended-31-december-2024.md0.65
  11. 2025-03-11Notice OF Results And Revised Trading Update2025-03-11_8773073_notice-of-results-and-revised-trading-update.md0.55
  12. 2025-01-29Trading Update For The Year Ended 31 December 20242025-01-29_8710458_trading-update-for-the-year-ended-31-december-2024.md0.55
  13. 2024-09-27Interim Results For The Six Months Ended 30 June2024-09-27_8443192_interim-results-for-the-six-months-ended-30-june.md0.58
  14. 2024-07-22New Licence Partner And Trading Update2024-07-22_8322026_new-licence-partner-and-trading-update.md0.38
  15. 2024-04-22Notice OF Agm Amp Publication OF Annual Report2024-04-22_8148276_notice-of-agm-amp-publication-of-annual-report.md0.43
  16. 2024-04-15Final Results For The Year Ended 31 December 20232024-04-15_8136376_final-results-for-the-year-ended-31-december-2023.md0.45
  17. 2024-04-09Updated Timing OF Full Year Results2024-04-09_8126986_updated-timing-of-full-year-results.md0.45
  18. 2024-03-14Updated Timing OF Full Year Results2024-03-14_8086752_updated-timing-of-full-year-results.md0.45
  19. 2024-01-24Trading Update2024-01-24_8003203_trading-update.md0.38
  20. 2024-01-18Investor Presentation2024-01-18_7994922_investor-presentation.md0.32
  21. 2023-11-30Trading Update2023-11-30_7914383_trading-update.md0.38
  22. 2023-09-28Interim Results2023-09-28_7782614_interim-results.md0.41
  23. 2023-07-25Trading Update2023-07-25_7652266_trading-update.md0.21
  24. 2023-03-24Final Results2023-03-24_7280385_final-results.md0.25
  25. 2023-01-24Trading Update2023-01-24_7499626_trading-update.md0.21
  26. 2022-09-22Interim Results2022-09-22_7419371_interim-results.md0.23
  27. 2022-04-08Notice OF Agm2022-04-08_6902064_notice-of-agm.md0.07
  28. 2022-03-17Final Results For The Year Ended 31 December 20212022-03-17_6968391_final-results-for-the-year-ended-31-december-2021.md0.25
  29. 2021-09-30Interim Results For 6 Months Ended 30 June 20212021-09-30_6598984_interim-results-for-6-months-ended-30-june-2021.md0.23

This research note was authored by a large language model after reading 25 regulatory filings published between 2021-09-30 and 2026-09-23. Each citation refers to a specific RNS announcement in the underlying data set. The note is an opinion, not advice. Do your own work before risking capital.