FuelCell Energy, Inc. (FCEL) | The Buildout — AI Infrastructure
The Verdict
FuelCell Energy designs and manufactures stationary molten carbonate fuel cell power plants. These modular systems run on natural gas, biogas, or hydrogen blends, delivering baseload electricity with integrated cooling capabilities. The company's technology targets data centers that face grid constraints and need reliable, on-site power.
| Market Cap | — |
| Revenue (TTM) | $168M |
| Revenue Growth | +29.7% |
| EBITDA Margin (TTM) | -69.1% |
| Net Cash | $214M |
| Earnings Beats | 4 of 7 |
| P/E (TTM) | — |
| EV/EBITDA (TTM) | — |
What We Like
- Pipeline of submitted proposals expanded to 4 GW, up more than 250% sequentially, with data center share rising to 89%.
- Proprietary 12.5 MW FuelCell Energy Block launched, purpose-built for data center behind-the-meter power with native DC output and absorption chilling that management says can reduce electric cooling load by 25–30%.
- Management reaffirmed its target of positive adjusted EBITDA once Torrington production reaches 100 MW annualized, up from current ~40 MW, and expects operating expenses to remain flat.
- Strong balance sheet with $441 million in cash and restricted cash as of April 2026, and additional non-dilutive financing from EXIM Bank ($49 million package approved).
- Technology has 50 years of cumulative utility-scale runtime across five installations, providing a track record that management believes de-risks adoption for hyperscalers.
What We’re Watching
- Conversion of the 4 GW pipeline into contracted backlog: management aims for FY2026 (by 31 Oct). Failure to deliver would severely damage credibility.
- ExxonMobil carbon capture demonstration: modules en route for June 2026 delivery, but the partner has not mentioned the project in its own earnings, raising questions about materiality.
- Legacy asset performance: the Groton Navy microgrid incurred a $42.6 million non-cash impairment; similar issues could emerge in other aging plants.
- Equity dilution: the company sold 15 million shares in and after Q2 FY2026, despite $441 million cash; continued dilution would pressure per-share value.
The thesis is strengthening in promise but remains unproven in delivery. The 4 GW pipeline explosion signals genuine customer interest, and the product offering is becoming more tailored to data centers. However, zero AI revenue and no binding contracts mean the investment case still rests entirely on future commercial conversion. The open question is whether management can convert proposals into orders before the fiscal year-end, as promised.
Earnings Beat
Revenue of $35.6 million declined 5% year over year, while adjusted EBITDA improved 12% to a loss of $17.1 million. Product revenue grew on Korean module deliveries, but the quarter included a $42.6 million non-cash impairment on the Groton Navy project, pushing the net loss to $78.7 million.
| Metric | Q2 FY2026 | Q1 FY2026 | Q2 FY2025 | YoY |
|---|---|---|---|---|
| Revenue | $36M | $30M | $37M | −4.8% |
| Gross margin | -36.3% | -19.2% | -25.2% | -1110bps |
| EBITDA | −$67M | −$16M | −$25M | +169.5% |
| EPS | $-1.43 | $-0.49 | $-1.75 | −18.2% |
| Pipeline proposals | 4.0 GW | >1.5 GW | n/a | — |
the world needs more power, clean, resilient, affordable, and continuous power—and that is exactly what we deliver.— Jason Few, President and CEO, Q2 FY2026 call, 8 Jun 2026
Management tone: Management struck a confident, promotional tone, centering the narrative on the AI opportunity. They were direct on operational details and cost discipline but reframed or avoided hard commitments on pipeline conversion timelines and future equity raises. The shift from prior quarters was palpable, with data centers now the dominant theme.
Management Guidance
Management reaffirmed its positive adjusted EBITDA target at 100 MW annualized production, without specifying a date. It committed to converting the 4 GW pipeline into contracted backlog by fiscal year-end 2026. Near-term product revenue is expected to be consistent from Korean module deliveries (GGE and CGN) in the second half of FY2026. The Torrington manufacturing expansion target was raised to 500 MW, executed in stages gated on backlog, at an estimated cost of $200–275 million. Capital raises remain an open possibility.
Trajectory
Revenue has grown 29.7% year over year on a trailing basis, driven by product deliveries to Korean customers, but gross margins remain deeply negative due to low production volumes and under-utilized capacity. Adjusted EBITDA losses have narrowed modestly as the company controls costs, though a large impairment charge pushed the bottom line deeper into the red. The path to profitability depends on reaching a 100 MW annual production run rate, 2.5x the current ~40 MW level, to achieve factory leverage.
The Model
The model projects FY+1 revenue of $143 million with an EBITDA loss of $66 million, improving to $250 million in revenue and breakeven EBITDA in FY+2. The near-term projection is anchored by existing service and generation backlogs plus Korean product deliveries, while FY+2 reflects the first contribution from potential data center order conversions.
| Metric | FY2025 | Next FY (E) | Following FY (E) |
|---|---|---|---|
| Revenue | $158M | $143M | $250M |
| YoY Growth | — | −9.6% | +74.8% |
| EBITDA | −$81M | −$66M | $0M |
| EBITDA Margin | -51.1% | -46.4% | 0.0% |
Projections are the median of 5 independent model runs.
Management reaffirmed its positive adjusted EBITDA target at 100 MW annualized production, without specifying a date. It committed to converting the 4 GW pipeline into contracted backlog by fiscal year-end 2026. Near-term product revenue is expected to be consistent from Korean module deliveries (GGE and CGN) in the second half of FY2026. The Torrington manufacturing expansion target was raised to 500 MW, executed in stages gated on backlog, at an estimated cost of $200–275 million. Capital raises remain an open possibility.
What Could Go Right — and Wrong
- Successful conversion of even a small fraction of the 4 GW pipeline into firm orders, triggering manufacturing expansion and revenue growth.
- ExxonMobil Rotterdam carbon capture pilot completes demonstration, generating commercial interest and a distinct second revenue stream.
- Fit Energy strategic agreement progresses to definitive phases, securing a large-scale anchor data center customer.
- Rapid ramp of Torrington production to 100 MW annualized, turning adjusted EBITDA positive ahead of the model's timeline.
- Standardized 12.5 MW block proves attractive, shortening sales cycles and improving product margins to the targeted 10–20%.
- No data center contracts signed by fiscal year-end, damaging management credibility and prolonging cash burn.
- Competitors like Bloom Energy win major hyperscale deals, establishing a reference base that FCEL lacks.
- Further legacy asset impairments, such as the Groton project, erode cash and distract from growth initiatives.
- Persistent equity dilution continues at current pace, significantly reducing total equity value for existing holders.
- Pipeline proves largely speculative, with few proposals advancing to contract, leaving revenue to rely on declining legacy streams.
Looking Ahead
The next twelve months are critical: management has committed to converting the 4 GW data center pipeline into contracted backlog by the end of fiscal 2026 (October). Nearer term, the ExxonMobil Rotterdam carbon capture modules are due for delivery in June 2026, with a demonstration later in the year. The Fit Energy agreement, signed in June, could progress to definitive phases. Manufacturing capacity expansions at Torrington will be gated on actual order flow, with the company targeting positive EBITDA at 100 MW annualized production. The narrative will be shaped by contract announcements—or their absence—over the coming quarters.
- June 2026ExxonMobil module delivery — Two carbon-capture modules arrive in Rotterdam; tests technology and partnership.
- 2H FY2026Korean module deliveries — Remaining GGE modules and start CGN deliveries; supports consistent product revenue.
- FY-end 2026Pipeline conversion deadline — Convert proposals to contracted backlog; crucial for AI credibility.
- TBD (post-24 Jun 2026)Fit Energy next steps — Up to 380 MW strategic agreement may advance to definitive purchases.
Financials
Annual Summary
| Metric | FY2025 | TTM |
|---|---|---|
| Revenue | $158M | $168M |
| Gross Margin | -18.9% | -18.2% |
| EBITDA | −$81M | −$220M |
| EBITDA Margin | -51.1% | -69.1% |
| Net Income | −$189M | −$223M |
| Free Cash Flow | −$148M | −$297M |
| Net Cash | — | — |
Key Ratios (Trailing)
- P/E TTM—
- EV/EBITDA TTM—
- EV/Revenue TTM—
- Price/FCF TTM—
- Gross Margin (TTM)-18.2%
- EBITDA Margin (TTM)-69.1%
- Net Margin (TTM)-132.8%
- ROIC-24.5%
- SBC / Revenue5.4%
The Company
FuelCell Energy manufactures stationary fuel cell power plants using its proprietary molten carbonate fuel cell technology. The systems are fuel-flexible, running on natural gas, biogas, or hydrogen blends, and provide continuous baseload power with low emissions. For AI data centers, the technology offers behind-the-meter generation with native DC output and integrated cooling, potentially reducing electric cooling load by 25–30%.
The company operates a 167,000-square-foot manufacturing facility in Torrington, Connecticut, currently running at an annualized capacity of ~40 MW, with plans to expand to 500 MW. It also has an R&D hub in Danbury, Connecticut, a solid-oxide engineering center in Calgary, Canada, and an assembly and service hub in Taufkirchen, Germany capable of up to 20 MW annually. FCEL sells through four business lines: Product, Service, Generation, and Advanced Technology, with revenue derived from a mix of product sales, long-term service agreements, electricity generation, and advanced technology contracts.
Business Segments
Competitive Landscape
FuelCell Energy competes against other fuel cell manufacturers like Bloom Energy and former licensee POSCO Energy, as well as conventional power providers such as GE Vernova, Caterpillar, and Cummins. The company argues its technology is differentiated by a 50-year utility-scale operating history, fuel flexibility, and integrated cooling for data centers. However, it lacks a large-scale data center reference installation, which may deter risk-averse hyperscalers.
- Bloom EnergyCompetes with solid oxide fuel cells for data center behind-the-meter power.
- POSCO EnergyFormer MCFC licensee; remains a competitor in Korean and international power generation markets.
- Fraunhofer IKTSOriginal technology licensor; could be a competitor now that FCEL's exclusive rights have ended.
- GE VernovaLarge-scale power generation competitor offering gas turbines for data center baseload.
- Caterpillar / CumminsDiesel and gas engine generators competing for distributed generation applications.
Supply Chain
FuelCell Energy relies on a limited number of suppliers for key components, with qualification taking 4–12 months. Its main manufacturing is in-house at Torrington, Connecticut. Disclosed customers are legacy utility and industrial plants, while data center engagement is still at the proposal stage.