FuelCell Energy (FCEL) Q3 2026 earnings review
Massive Data Center Wins Overshadowed by Margin Collapse and Dilution
FuelCell Energy successfully pivoted its narrative to AI/Data Centers, booking a massive up to 380 MW deal with Fit Energy and driving 'Awarded Backlog' up by $2.35B. However, Q3 revenue fell 29% YoY to $33.0M, and the gross loss exploded to $(24.5)M. The primary reason contradicts the bullish sales narrative: FCEL recorded a $17M charge because the contractual pricing for the highly touted Fit Energy Phase 0 deal is below current manufacturing costs. Meanwhile, the company's robust $737M cash pile was bought at the cost of relentless shareholder dilution, with outstanding shares nearly doubling over the last nine months.
🐂 Bull Case
The up to 380 MW agreement with Fit Energy and a new 75 MW Capacity Reservation Agreement in Texas prove FCEL can bypass grid bottlenecks and win utility-scale data center workloads.
With $737.3M in total cash, the company has fully funded its Torrington facility capacity expansion to 500 MW without needing immediate external capital.
🐻 Bear Case
The $17.0M inventory valuation charge proves the initial 30 MW Phase 0 of the Fit Energy deal was priced below current manufacturing costs, calling long-term unit economics into question.
EPS 'improved' to $(0.64) from $(3.78) YoY, but this is an illusion driven by a massive surge in share count (outstanding shares jumped from ~46M in Oct 2025 to ~80M in July 2026) due to heavy ATM and underwritten offerings.
⚖️ Verdict: 🔴
Bearish. The commercial momentum in the data center market is undeniably real, but winning deals by pricing below current cost while heavily diluting shareholders is a destructive financial combination in the near term.
Key Themes
The True Cost of Data Center Wins
Management heralds the Fit Energy Capital Equipment Purchase Agreement (CEPA) as a major commercial milestone. However, the data contradicts the narrative of profitable growth: FCEL absorbed a $17.0M charge in Q3 because product costs and fixed manufacturing overhead exceed the contractual pricing established for the 30 MW Phase 0. While management expects scaling to 100+ MW to eventually lower per-unit costs below this price point, they are currently buying market share at a severe margin penalty.
AI & Data Center Pipeline Explosion
The primary growth driver is accelerating power demand from AI workloads. The sales pipeline surged to 10 GW in Q3 (up from 4.4 GW in Q2), with high-density data centers making up 97% of proposals. The platform's native DC power backbone and ability to bypass 5-7 year grid interconnection delays are strongly resonating with hyperscalers.
Relentless Shareholder Dilution
FCEL ended the quarter with an impressive $737.3M in cash, but the cost to shareholders was staggering. The company sold 12.3M shares in a July public offering (raising $245.5M) and another 4.1M shares via its ATM program (raising $52.9M). Total outstanding shares surged from 46.1M at FYE 2025 to 80.0M by Q3 2026. The 'improving' EPS is a direct mathematical consequence of share bloat, not bottom-line fundamental improvement.
Carbon Capture Progressing to Demonstration
A significant technology validation occurred as FCEL delivered its first two carbon capture modules to ExxonMobil's Rotterdam facility. Unlike conventional systems, FCEL's carbonate fuel cells capture CO2 while simultaneously generating power and hydrogen. Successful industrial-scale demonstration here is required to unlock a massive point-source emission market.
Generation Portfolio Offline Drag
Generation revenue decelerated 29% YoY to $8.8M. The drop was significantly exacerbated by the 7.4 MW Groton Project at the U.S. Navy base, which remained offline pending a previously announced equipment upgrade (to three standard 2.5 MW blocks). The timeline for its return to service remains unclear, removing a chunk of high-margin recurring revenue.
Profitability Goalposts Pushed Out
Management shifted its target for achieving positive Adjusted EBITDA to Q4 FY2027. This is highly conditional: it requires converting the newly 'Awarded' backlog into 'Committed' backlog, executing customer deliveries flawlessly, and ramping the Torrington annualized production rate from its current ~37.1 MW up to 100 MW by October 2026.
Other KPIs
Decelerating significantly, down 30% YoY from $26.0M. The decline was primarily due to fewer module deliveries to customers in South Korea compared to the prior year, offsetting the early momentum from domestic data center awards.
Deteriorating from $(16.4)M in the prior year quarter. The 124% expansion in Adjusted EBITDA losses strips away the excuse of non-cash impairments and lays bare the severe impact of the $17.0M inventory valuation charge on the Fit Energy CEPA.
Guidance
Accelerating. The company is currently operating at an annualized rate of 37.1 MW and plans a steep ramp over the next quarter to hit 100 MW, which management claims is the threshold required to eventually reach positive Adjusted EBITDA.
Stable. The company confirmed its long-term physical capacity expansion is fully funded and on track, progressing with tape caster installation and factory design engineering.
Decelerating / Pushed Out. Management explicitly pegged this timeline to the successful conversion of Awarded Capacity to Committed Backlog and the execution of cost reduction initiatives.
Key Questions
Unit Economics of Fit Energy Phases 1-3
You took a $17M charge on Phase 0 of the Fit Energy CEPA because pricing is below current cost. Are Phases 1 through 3 locked into this exact same pricing structure, and if so, what precise production volume must Torrington achieve to ensure those future phases do not also generate negative gross margins?
Dilution and Capital Strategy
With a record $737M in cash after significant equity issuances, is the company now fully funded to reach the 500 MW capacity and Q4 2027 profitability targets, or should investors anticipate continued usage of the ATM program?
Groton Project Timeline
The Groton Project being offline significantly pressured Generation margins. What is the specific timeline for the completion of the 2.5 MW block upgrades, and when will it resume contributing to recurring generation revenue?
Awarded vs Committed Conversion
Of the $2.35B in new 'Awarded Capacity Backlog', what are the specific triggers or milestones that Fit Energy must clear to exercise their options for Phases 1, 2, and 3, and convert this into firm Committed Backlog?