ITM Power

Phase A — Understand the business

Company Overview

The model. ITM designs and manufactures PEM electrolyser stacks and plants. Renewable electricity + water in, green hydrogen out. Revenue is overwhelmingly equipment sales — H1 FY2026 revenue of £18.0m split £15.5m equipment / £2.5m engineering, spares and maintenance. It is a lumpy, project-driven, contract-manufacturing P&L, not a recurring-revenue business — today.

Product ladder (2026). Built on the current TRIDENT PEM stack (~670 kW/stack):

• NEPTUNE — 2–5 MW plug-and-play plant for small/mid projects (NEPTUNE V = 5 MW).
• POSEIDON — 20 MW skid-mounted, pre-fabricated/pre-tested module for large projects.
• ALPHA / ALPHA Fifty — 50 MW prefabricated plant launched Oct 2025 at ~€50m list.
• CHRONOS — the next-gen stack (2–2.5 MW/stack vs Trident's 670 kW), claimed −40% manufacturing cost, +10% efficiency, −40% iridium, >50% fewer parts; COD targeted 2028. This is the roadmap that the whole bull case leans on.
• Hydropulse (launched Jun 2025) — a build-own-operate subsidiary (Hydropulse GmbH, Germany) selling "green hydrogen as a service" under long-term offtake, using NEPTUNE plants. CEO Schulz: molecules + services will "eventually overtake" electrolyser sales as the leading profit source. Strategically the most important — and most capital-hungry — new leg.

Customers. Blue-chip but concentrated and FID-dependent: Shell (REFHYNE II, Rhineland — 100 MW Trident reservation), RWE (150 MW), Yara (24 MW Herøya green-ammonia plant, live), Uniper (Humber H2ub, 6× POSEIDON = 120 MW), Octopus Energy and MorGen Energy (20 MW, Wales) in the UK, plus Norway's Agder Hydrogen Hub (4× NEPTUNE V). our figures is empty — all.

Contract structure — the crucial nuance. The firm order backlog is £152m (H1 FY2026), of which 71% is now profitable and 29% legacy loss-making, to unwind over ~18 months. The eye-catching 710 MW "Stablegrid", 150 MW RWE and 100 MW REFHYNE II figures are capacity-reservation agreements, not firm backlog — options contingent on the customer's final investment decision (FID). The gap between the MW headline pipeline and the modest £-backlog is the single most important thing to hold in mind for every later lens.

History. Founded June 2000 in Saffron Walden, Essex (Highgate/Lloyd/Wreford), originally a fuel-cell company; AIM IPO 2004 raising £10m. HQ now Sheffield (Bessemer Park gigafactory + two further Sheffield sites) plus an office in Hesse, Germany.

Supply Chain

Upstream inputs → ITM → end customer, named where sourced (supply-chain.md missing — all ):

• Iridium & platinum (anode/cathode catalysts) → catalyst suppliers (e.g. Heraeus, Johnson Matthey historically) → ITM stacks. This is the chokepoint. Iridium is produced only as a by-product of platinum mining, ~7.5 tonnes/yr globally; PEM needs ~400 kg Ir per GW and iridium is ~20–25% of stack cost; credible shortage risk by ~2030 if PEM scales. South African/PGM mining (Anglo American Platform, Sibanye, Impala) is the ultimate single-region dependency.
• PFSA membrane (Nafion-type) → Chemours / Gore / AGC — a second concentrated input; PFAS regulatory pressure in the EU is a latent risk.
• Titanium PTLs, bipolar plates, balance-of-plant (rectifiers, gas separators, compressors) → European industrial supply chain.
• Manufacturing → Sheffield Bessemer Park gigafactory (nameplate ~1 GW/yr; AutoStacker robotic assembly line rated >2 GW/yr of stacks).
• EPC / integration → ITM Linde Electrolysis (ILE), the 50/50 JV with Linde Engineering, wraps ITM stacks into turnkey ≥10 MW plants (Leuna, REFHYNE).
• End customers → refiners, ammonia/fertiliser (Yara), utilities (RWE, Uniper, Octopus), and now defence-adjacent PtX (Rheinmetall).

Chokepoints: (1) iridium — structural, industry-wide, and the reason CHRONOS's −40% iridium claim matters; (2) the Linde relationship — Linde is simultaneously ~16% shareholder, JV partner, EPC channel and a customer, which concentrates both dependency and conflict-of-interest surface.

Competitive Advantages (moats)

Technology position. PEM's genuine advantages over alkaline are fast dynamic response (ramps with intermittent renewables), high current density and small footprint — real for grid-coupled green-H2. ITM is one of the few PEM pure-plays at scale (peers: Plug Power, Siemens Energy Silyzer, Cummins/Accelera, Ohmium; Nel does both).

But the moat is thin. (1) No cost moat — alkaline (thyssenkrupp nucera, John Cockerill, LONGi, Chinese OEMs) is structurally cheaper on nickel/iron and dominates GW-scale tenders; Chinese alkaline is a deflationary force. (2) Switching costs are project-level, not franchise-level — each plant is re-tendered. (3) IP is contestable — PEM know-how is diffusing; iridium-reduction is an industry race, not an ITM-only breakthrough. (4) Bargaining power is weak — a sub-£45m-revenue supplier negotiating with Shell/RWE/Uniper has little pricing leverage, which is exactly why the legacy book was loss-making.

Where a real edge could form: (a) CHRONOS + AutoStacker automation if the −40% cost claim survives contact with production — a manufacturing/cost edge is the only durable moat in commoditising hardware; (b) Hydropulse converting ITM from price-taking vendor to offtake-owning producer, which would create recurring revenue and demand-pull it controls; (c) UK-sovereign / NATO positioning (GB Energy stake + Rheinmetall) as a non-Chinese, allied-supply-chain electrolyser champion. All three are promises, not moats, in July 2026.

Segments

our figures empty → all. ITM does not report clean product-segment P&L; the meaningful cuts are revenue line and geography:

• By revenue line (H1 FY2026): Equipment £15.5m / Engineering-spares-maintenance £2.5m. The aftermarket/services line is tiny today but is the seed of the Hydropulse "molecules & services" pivot.
• By geography: Germany is now the primary commercial market — the 710 MW Stablegrid reservation, 150 MW RWE, 100 MW REFHYNE II and the Hydropulse GmbH BOO vehicle are all German. UK second (Uniper Humber, MorGen, Octopus, and the GB Energy/CHRONOS manufacturing base). Norway third (Yara, Agder).
• Trend & cause: Revenue is accelerating fast off a tiny base — £26.0m FY2025 (+~50% YoY), guided £40–43m FY2026 (+~35–55%), "~400% over two years / ~600% over three". The driver is the Schulz reset: fewer standardised platforms (NEPTUNE/POSEIDON/ALPHA), disciplined bidding, and the legacy loss-book rolling off. The mix is shifting from loss-making legacy to profitable standardised contracts — the most important operational trend in the whole file.

---

Phase B — Measure performance

Earnings Result (latest print + trajectory)

Latest reported: H1 FY2026 (six months to 31 Oct 2025), reported 29 Jan 2026:

• Revenue £18.0m — record half-year (vs £15.5m H1 FY2025).
• Gross loss £6.5m — improved from £10.2m.
• Adjusted EBITDA loss £11.9m — narrowed sharply from £16.8m.
• Cash £197.8m, down only ~£5m YoY (from £203.1m at 31 Oct 2024) — near-flat despite the EBITDA loss, because a growing order book brings customer advance payments (working-capital inflow).
• Order backlog £152m, 71% profitable (up from 60% at Apr 2025).

FY2026 guidance (reiterated Jan 2026, upgraded in the ~May 2026 year-end trading update): revenue £40–43m (from £35–40m; +35% YoY), adjusted EBITDA loss £27–29m (£4m YoY improvement, implying FY2025 adj EBITDA loss ~£31–33m ), year-end cash £170–175m (pre the Q2-2026 GB Energy raise). The May update said FY revenue would land ahead of guidance with EBITDA loss within range.

Balance sheet. Net cash ~£175–215m depending on the timing of the £40m Great British Energy subscription (Q2 2026); zero debt. This is the opposite of a going-concern story — ITM is cash-rich; the risk is not insolvency but value-destructive cash deployment.

Market reaction / read-through. The tape has treated each print as confirmation of the turnaround; combined with the Rheinmetall LOI, GB Energy stake and MSCI UK Small Cap inclusion, the stock re-rated +~260% over the trailing year. Unusual vs its own history: losses are narrowing on rising revenue and cash is holding — genuinely better than the 2022–23 cash-burn-and-warnings era. The concern is entirely valuation and demand-conversion, not the print itself.

Earnings Calls (sentiment trend)

From the H1 FY2026 call (29 Jan 2026):

• Tone: confident, disciplined, patient. Management repeatedly frames the story as execution over hype — a deliberate contrast with the pre-2022 promotional era. CEO Schulz: "We need to really focus on winning, and then executing profitable contracts… maintain the discipline on overheads and cost control."
• What they now emphasise (that they did not in 2021): profitable-contract %, cost discipline, cash preservation, standardised products, revenue-recognition rigour (a shift to percentage-of-completion for POSEIDON/ALPHA/custom NEPTUNE, improving visibility).
• New strategic drumbeat: "Hydropulse will buy the electrolysers… and operate plants to supply hydrogen to industrial customers under long-term offtake contracts" (Schulz) and the CFO's "the potential… to grow shareholder value is vast." The CTO ("Simon") called the next-gen stack (rendered "KRONOS" on the call; officially CHRONOS) "a genuine game changer."
• Sentiment shift over time: 2021 = growth/capacity euphoria (5 GW ambitions); 2022–23 = contrition and reset ("unacceptable performance," exit hydrogen refuelling, staff cuts); 2024–26 = quiet-confidence execution narrative. The arc is credible and consistent — the tone risk now is the opposite of 2021: management talking up Hydropulse/CHRONOS/defence into a re-rated price.

Comps

PEM/electrolyser and adjacent clean-hydrogen names. Market caps with dates; multiples not sourced this run are n/a — none are fabricated.

Read: ITM at ~£800m/~14x forward sales is richly valued for a sub-£45m-revenue, lossmaking hardware maker, and sits above alkaline leader thyssenkrupp nucera on market capMarket capitalisationThe share price multiplied by the number of shares. What the market says the equity is worth. despite nucera having real revenue and the cheaper, GW-dominant technology. The peer set is uniformly pre-profit or licence-model; there is no clean earnings multiple to anchor to — which is precisely why valuation is contested (see Lens 12).

Stock-Price Catalysts (>5% moves, ~5 yrs)

The name is a volatility machine — a textbook thematic boom-bust-boom:

• 2020–Jan 2021 — the hydrogen bubble. Ran to an all-time high ~700p (peak ~26 Jan 2021); £250m raise (Oct 2021) to fund a since-abandoned 5 GW capacity plan.
• 2022 — three profit warnings in ~8 months. Manufacturing/tooling delays on new stacks, legacy-contract losses, warranty provisions, inventory write-downs, Leuna delays. The 27 Oct 2022 unscheduled trading update drove shares −37% (104p→66p intraday) — later the subject of an FCA insider-dealing case (Lens 10).
• 2022–2025 — the bust and reset. Graham Cooley out; Dennis Schulz in (Dec 2022); exit hydrogen refuelling, staff cuts, standardise products. Trough ~57p (52-wk low).
• 2025–2026 — the re-rating (+~260% YoY). Record revenue/backlog prints, MSCI UK Small Cap inclusion, £46.5m DESNZ CHRONOS grant + £40m Great British Energy ~10% equity stake, and the April 2026 Rheinmetall NATO-PtX LOI (which spiked the stock). Ran to a 52-wk high ~217p, currently ~113p.

What the market actually reacts to: (1) the hydrogen theme (sector risk-on/off dominates the beta), (2) government money and index flows, (3) marquee partner headlines (Shell, GB Energy, Rheinmetall) — far more than fundamentals. Order conversion and cash discipline have mattered less to the tape than narrative catalysts. That cuts both ways for anyone taking a position.

---

Phase C — Judge people & books

Management

• CEO — Dennis Schulz (since 1 Dec 2022; ex-Linde Engineering). Track record: the credible architect of the turnaround — narrowed losses, lifted profitable-contract mix 60%→71%, held cash near-flat, killed the loss-making refuelling arm, standardised the product line, and launched two genuinely strategic initiatives (Hydropulse, CHRONOS). His Linde pedigree matters: it aligns ITM with its largest strategic shareholder and its EPC channel. Archetype: professional operator/turnaround manager (not founder) — appropriate for this stage. Watch: he is now visibly selling a Hydropulse/CHRONOS/defence narrative into a re-rated price; the incentive to over-promise has returned.
• CFO — Amy Grey (joined 2025; predecessor Andy Allen moved to VP Strategy Jan 2025).
• CTO — "Simon" (first-named on the Jan-2026 call; long-tenure stack engineer) — owns the CHRONOS roadmap.
• Chair — Sir Roger Bone retired 15 Jan 2026 (Chair since 2019); a permanent successor was not sourced this run. Board upgraded: Sir Warren East (ex-ARM CEO 2001–13, ex-Rolls-Royce CEO 2015–22) and John Howarth joined as NEDs post the Oct-2025 AGM; Oleg Williamson (Linde) joined as NED Jun 2026. Warren East's arrival is a real quality signal — a semiconductor/precision-engineering heavyweight on a manufacturing-cost-curve story.
• Capital-allocation history — mixed-to-improving. The 2021 £250m raise funded a 5 GW plan that was scrapped, and the pre-2022 book destroyed value on underpriced contracts — a poor record. Under Schulz, allocation is disciplined (right-sized capacity to 1.5 GW, cash preserved). The open question is Hydropulse: owning BOO plants is capital-intensive and would consume the net-cash cushion — the next allocation test.
• Skin in the game / ownership: ~16.2% "not in public hands"; Linde ~16.2% (20% in 2019 → 17.25% in 2020 → 16.2% Aug 2024) and Great British Energy ~10% are the anchors. our figures absent; a CFO on-market purchase was noted.
• Red flags: the Linde web (shareholder + JV partner + EPC + customer + board seat) is a related-party concentration to monitor; and management's renewed promotional tilt at a high price.

Forensic Red Flags

our figures empty; no SEC filings (no CIK). Assessment from public results + press:

• Revenue recognition — the single most important accounting item. ITM is moving from completed-contract to percentage-of-completion for POSEIDON/ALPHA/custom NEPTUNE. POC accelerates revenue and injects management estimation into the top line — legitimate for long-cycle equipment, but exactly the lever that flatters a growth story. Watch for contract-loss provisions and margin restatements as legacy work completes.
• Legacy-contract & warranty provisions — 29% of backlog is still legacy/loss-making; the 2022 blow-up was warranty + inventory write-downs. Provisioning adequacy is the number-one thing to audit in the FY2026 accounts.
• Working-capital optics — near-flat cash despite EBITDA losses is driven by customer advances; healthy while orders grow, but it reverses if order intake stalls (advances unwind into cash burnCash burnHow much more cash goes out than comes in, per period. The clock on a company with no profits.). Do not read flat cash as underlying profitability.
• Adjusted vs statutory — the headline is adjusted EBITDA; SBC, exceptionals and capitalised development costs sit outside it. Reconcile to statutory operating loss and to cash.
• Grants & equity — the £46.5m DESNZ grant and £40m GB Energy equity are non-operating inflows; strip them out to see true operating burn.

Regulatory findings (required).

• SEC (EDGAR LR/AAER): none possible — no CIK; ITM is not an SEC filer.
• UK FCA: no enforcement against the company. The FCA banned and fined an individual — Neil Dwane, a Capital Markets Adviser in ITM's Investor Relations team — £100,281 for insider dealing ahead of the 27 Oct 2022 trading update (he sold 125,000 shares the day before the −37% fall, then repurchased lower). This reflects on an individual, not on ITM's accounts, but it does confirm the materiality of the Oct-2022 information event.
• FRC / restatement / class action: none surfaced.
• Net: No material corporate regulatory or accounting-enforcement findings — verified via the absence of an SEC CIK, UK FCA/FRC web search, and public results, as of 2026-07-10. The one enforcement action is against a former IR individual, not the company.

---

Phase D — Project & stress-test

Forward Projection

Bottom-up from FY2026 guidance (rev £40–43m, adj EBITDA loss £27–29m, net cash ~£175–215m). ITM is lossmaking, so the honest projection is revenue + EBITDA path + net-cash runway, not EPS (EPS stays negative across the window; a P/E is meaningless). All, inputs labelled.

Inputs: base assumes reservation-to-order conversion of 1–2 marquee German projects/yr, standardised-product gross margin turning positive by FY2027, CHRONOS cost-down from 2028, and Hydropulse contributing first molecule revenue FY2028+. Runway is not the risk (£190m net cash vs ~£25–29m annual EBITDA loss ≈ 6–7 years of runway even before grants ) — demand conversion is.

Per --watchlist rules and the brief, no our model create is logged (breadth mode). The scoreable base-case forecast to log later, on a our position log pass: "ITM.L books ≥200 MW of firm new electrolyser orders (not reservations) by FY2027 year-end" and "ITM.L reaches adjusted-EBITDA breakeven by FY2028" — the two hinges of the whole thesis.

Bull vs Bear

Bull case. A genuinely de-risked turnaround with an option on the entire European green-H2 build-out. Losses halving, 71% profitable backlog, ~£190m net cash and zero debt, a right-sized 1.5 GW factory, a next-gen stack (CHRONOS) targeting −40% cost / −40% iridium, and a demand-creation engine (Hydropulse) that could convert ITM from lumpy vendor to recurring hydrogen producer. Sovereign/allied tailwinds — GB Energy's ~10% strategic stake and the Rheinmetall NATO-PtX LOI — position ITM as the non-Chinese, UK-based electrolyser champion at the exact moment energy security is a policy priority. If even a fraction of the 710 MW Stablegrid / 150 MW RWE / 100 MW REFHYNE II reservations convert to firm orders, revenue compounds off a tiny base and the ~14x sales looks cheap in hindsight. Berenberg's 200p (Buy) reflects this path.

Bear case (2–3 permanent-impairment risks). (1) The demand isn't there. The green-H2 market is in a FID graveyard — the EU Hydrogen Bank's 2nd auction (late 2025) saw 7 projects / 1.88 GW of 2.33 GW winning capacity withdraw despite €1.2bn of grants, and US 45V was gutted by the OBBBA (construction-start cutoff ~2027–28); ~50+ projects publicly cancelled; unsubsidised green H2 is $2.50–7.00/kg and uncompetitive. ITM's reservation pipeline is options on projects that keep dying. (2) No cost moat vs cheap Chinese/alkaline — GW-scale tenders favour alkaline economics; PEM's iridium exposure is a structural cost headwind CHRONOS only partially offsets. (3) Hydropulse turns a cash-rich equipment maker into a capital-intensive project developer competing with its own customers and burning the net-cash cushion. Pre-mortem (18 months out, thesis broken): two flagship German FIDs slip to 2028+, order intake stalls, customer advances unwind into real cash burn, the Rheinmetall LOI never becomes a firm order, the hydrogen theme rolls over, and the stock round-trips toward its ~57p low / net-cash floor (~27p/share) — a ~40–75% drawdown from ~113p with the business itself unharmed.

Are multiples too high? On any fundamental basis, yes — ~14x forward sales for a lossmaking hardware maker above a revenue-generating alkaline leader is a theme/sovereign premium, not an earnings-supported valuation. Contrarian view (what the market refuses to see): the market is treating the Rheinmetall/GB-Energy/CHRONOS narrative as demand proof when it is demand hope — the one metric that would settle the debate, firm-order conversion of the MW reservations, has not yet happened, and the broader industry is providing daily evidence that European hydrogen FIDs are the binding constraint, not electrolyser supply.

Devil's Advocate (short-seller)

Dismantling the bull case. Where the money is concentrated: a handful of German mega-projects (Stablegrid, RWE, REFHYNE II) that are reservations, not orders — if two slip or die, the growth story evaporates. Why the moat is weaker than bulls think: ITM sells a commoditising machine into re-tendered projects with no pricing power against Shell/RWE/Uniper; the cheaper technology (alkaline) wins GW-scale, and China is the deflation engine. Most dangerous competitor bulls underestimate: Chinese alkaline OEMs (LONGi, and the CJH/Cockerill-Jingli axis) on price, plus thyssenkrupp nucera on scale and balance sheet. Worst capital-allocation risk: Hydropulse — funding build-own-operate plants would convert the pristine ~£190m net-cash balance sheet (the actual reason to own the stock through the last bust) into an illiquid, capital-hungry developer book, and puts ITM in competition with its own electrolyser customers. Assumptions that must hold for ~113p: European hydrogen FIDs re-accelerate, reservations convert, CHRONOS hits its −40% cost target on schedule (2028), and the theme premium persists. If growth disappoints 20–30%: at ~14x sales the de-rating is violent — the floor is net cash (~27p/share), and the stock has visited ~57p within the last year. Single permanent-impairment scenario: a second wave of legacy/warranty provisions on POC-recognised contracts (a repeat of 2022) that shatters the "we've fixed execution" narrative — plausibility low-to-moderate, but it is the exact failure mode this company has already lived through once.

Management Questions (ordered by information value)

1. Of the £152m firm backlog and the ~960 MW of headline reservations (Stablegrid 710 / RWE 150 / REFHYNE II 100), how much MW has a signed, unconditional purchase order — and how much is contingent on a customer FID that has not yet been taken?
2. What is the firm-order intake run-rate you need per year to hit adjusted-EBITDA breakeven, and are you currently above or below it?
3. How much capital will Hydropulse consume over the next three years, where does it come from (own balance sheet vs project finance vs partners), and at what point does BOO capex draw down net cash?
4. Does Hydropulse compete with your electrolyser customers, and how do you manage that channel conflict?
5. On CHRONOS: is the −40% manufacturing cost and −40% iridium validated in production hardware or still a design target, and what de-risks the 2028 COD?
6. What are your gross-margin bridge assumptions to sustainably positive product gross margin — pricing, standardisation, volume — and when?
7. How exposed is your cost curve to iridium price/supply, and what is your hedging / thrifting / recycling strategy this decade?
8. How do you win GW-scale tenders against alkaline (thyssenkrupp nucera, Chinese OEMs) on cost — or do you concede that segment and focus on grid-coupled PEM niches?
9. Under percentage-of-completion, what governance prevents a repeat of the 2022 contract-loss / warranty surprises?
10. What are the milestones and clawbacks on the £46.5m DESNZ grant, and what does GB Energy's ~10% stake entitle it to (board, IP, procurement)?
11. Is the Rheinmetall LOI progressing toward a firm order, and what volume/timeline would make Giga-PtX material?
12. How should we think about Linde as simultaneously your largest strategic holder, JV partner, EPC channel, board member and customer — where are the related-party guardrails?
13. What is your capacity-utilisation today against the 1.5 GW nameplate and >2 GW AutoStacker rating, and at what utilisation do fixed costs stop dragging gross margin?
14. What order-book or cash trigger would make you pause Hydropulse or CHRONOS capex to protect the balance sheet?
15. Given the ~260% re-rating, what do you think the market is pricing in that you would caution against?

---