PowerCell Pivots to Water and Blue Hydrogen — Right as the Tape Abandons AI Data Centers

Revenue falls 61%, cash falls from SEK 76m to SEK 18m, and the growth story quietly migrates from green hydrogen to natural gas reforming and data-center co-products.

PCELL.ST · Earnings Call · 2026-10-22

PowerCell Pivots to Water and Blue Hydrogen — Right as the Tape Abandons AI Data Centers

A quarter the company says it prepared for, and one it clearly didn't enjoy

PowerCell's third quarter is a small-cap industrial in full defensive posture. “Revenue declined 61%.” — Richard Berkling, Executive (likely CEO or similar) · 2026-10-22 Management attributes the hole to two separable timing effects: supply-chain slippage on the PS190 ramp that pushed secured Power Generation deliveries out of Q3, and automotive royalties that failed to arrive because Chinese deployment ran far below plan — Richard Berkling describes an industry that expected growth in 2026 and instead saw a market “with up to 55%, 60%” — Richard Berkling, Executive (likely CEO or similar) · 2026-10-22 decline. Anders During's framing is blunt and oddly valedictory: “the numbers are not the ones that we want to report to you, but they are the numbers that we are prepared for.” — Richard Berkling, Executive (likely CEO or similar) · 2026-10-22 The cost side is where the company has actually delivered. Combined selling, administration and R&D fell 25% year over year in the quarter, and the programme is not finished — management points to a cost reduction of roughly SEK 50 million off the annual cost base by 2027 versus 2025. Year-to-date operating cash flow is negative SEK 92 million, though the CFO's rolling-twelve-month framing — that the company now hovers around zero in a very tough situation — is the more charitable read, and arguably the fairer one.

What actually changed: the molecule, and the by-product

The most interesting thing in this transcript is not the revenue line. It is a wholesale re-sorting of what PowerCell says its technology is worth. Compare the keyword record: Green hydrogen was the company's single most prominent theme in the prior quarter's transcript. This quarter it has disappeared from the top thirty entirely. So have gray hydrogen and hydrogen availability as standalone themes. In their place, blue hydrogen climbs into the top ten, natural gas becomes a headline concept, and the two biggest new entries of the quarter are water and megawatt. This is not a rhetorical flourish. Berkling spends an unusual amount of the prepared remarks arguing that natural gas reformed into hydrogen with carbon capture — at indicative cost below $2 per kilogram in several U.S. projects — is the thing that unlocks scale, because "hydrogen supply determines which project can move forward and at what economics." A year ago on this same call, he acknowledged the tension directly: “this contradicts a bit PowerCell's background as a very green company” — Richard Berkling, CEO · 2026-07-16 — before making exactly the same argument. The company has now decided to stop apologising for it. The water angle is stranger and, frankly, more original. Berkling argues the by-product is a permitting asset in water-stressed data-center regions:

if we have a 50-megawatt fuel cell installation running in a data center and the average load is half, 25 megawatts, that means that we produce more than 250 tonnes of fresh water per day... the asset of water production has been massively underestimated by us before.
Richard Berkling, Executive (likely CEO or similar) · 2026-10-22

That is a genuinely new claim in this company's disclosure history, and it comes tied to a new commercial concept — energy orchestration — plus a new entry in the trajectory, power utility, which did not appear in any prior quarter's top thirty and now ranks near the top of the list. PowerCell is repositioning from a fuel-cell vendor to a seller of energy services into utilities and hyperscalers.

Data centers aren't new. The role is.

Careful reading matters here. Data centers have been discussed on this call series since at least October 2024 and were the company's top keyword in Q3 2025. In that earlier call the pitch was explicitly complementary: PEM beats solid-oxide on dynamics, so PowerCell would take backup power and peak shaving while Bloom Energy kept the baseload. Today the claim is much larger — large installations of 50 megawatts and above, prime power, and a 3-megawatt forty-foot container holding 16 to 20 PS190 systems against an industry norm of 1 to 1.2 megawatts. The supporting evidence is thin but real: an ECL order in July that provided the first multi-megawatt reference for AI data centers, a small DLR order demonstrating fuel-to-power, and yesterday's non-binding MOU with Bosch covering joint development and execution of 50MW-plus installations in the U.S., Canada, Europe and the Middle East. The collaboration with Bosch is expected to throw off roughly EUR 5 million of revenue in Q4 — a number doing disproportionate work in this quarter's story. And here is the tension. The equity tape has spent the last ninety days de-rating precisely this cluster. AI data center exposure is down about 5.6% over 90 days across a 54-name basket with six advancers against forty-eight decliners; data center power is down 5.1%, data center capacity 8.0%, high-bandwidth memory 6.9%. Even the 360-day window that was kind to data-center interconnect has rolled over at the short end. PowerCell is accelerating its narrative into the one theme the market has been quietly marking down since midsummer. Bloom Energy, its named competitor, reported in the same window leaning on grid power, electricity demand and 800V DC — same argument, deeper pockets, and a share price that has already been voted on.

Cash: from comfortable to contingent

In July, with SEK 76 million on the balance sheet, the CFO's answer on liquidity was: “We tackle the problem by being careful about how we spend cash... We feel comfortable at the level of where we are.” — Anders Düring, CFO · 2026-07-16 Three months and one 61% revenue decline later, the answer has changed shape: “we have SEK 18 million in cash. However, we also have an unutilized check credit of SEK 50 million.” — Anders During, Executive (likely CFO or similar) · 2026-10-22 That is a 76% sequential decline in the headline number, papered over by drawing capacity. The SEK 68 million combined figure is what management says carries the company forward. The sequencing matters. The EUR 5 million Bosch payment, the Q4 catch-up deliveries, and the remaining cost actions all have to land roughly on time for that arithmetic to stay comfortable — and the MOU itself is unsigned. Berkling is unusually explicit about the asymmetry: “until anything is signed, nothing is signed.” — Richard Berkling, Executive (likely CEO or similar) · 2026-10-22 Against a cautious outlook carried into 2027 as well, that is not a buffer. It is a schedule. The read: PowerCell has done the hard, unglamorous thing — cutting the cost base into a shrinking business — and has simultaneously found a genuinely differentiated story about water, density and blue hydrogen. But it is telling that story into a tape that has stopped paying for data-center exposure, with SEK 18 million of cash and an unsigned MOU. The company is more interesting than it was six months ago. It is also less insured.