Report contents
Industry landscape
Where the company competes: a project-by-project fight with the local grid
Bloom Energy makes solid oxide fuel cells and packages them into on-site generators it calls Energy Servers. A single module produces roughly 200 to 300 kilowatts of baseload electricity from natural gas, biogas or a hydrogen blend, and customers group them into multi-megawatt installations. The company's 10-K splits revenue into product and installation, service, and electricity sold under long-term contracts. That three-way split is the best guide to the market: Bloom does not really sell into a "fuel cell industry." It sells on-site baseload power to facilities that cannot tolerate outages or that pay high retail rates, including data centers, hospitals, semiconductor fabs, retailers and, in Korea, utility-scale buyers.
The value chain upstream is unremarkable. The stacks use ceramics, stainless steel, nickel and small amounts of rare earths; none of those suppliers holds pricing power over Bloom, and cell manufacturing is the piece of the chain that captures the technology rent. The binding upstream input is the fuel itself. A Bloom installation is only as good as the gas line behind it, and gas price and availability at the site sit outside the company's control. Downstream sits the customer's electricity bill, and there the reference price is the local utility's retail rate. Bloom earns a margin only when its levelized cost of electricity (the all-in cost per kilowatt-hour across fuel, capital and service over the contract term) undercuts what the grid charges for the same reliability. The party with real bargaining power is therefore neither Bloom nor a rival fuel-cell maker; it is the utility that sets the price to beat, together with the regulators who fix tariffs and set the investment tax credit that improves Bloom's payback.
That price-to-beat is set locally, so the competitive arena is the individual site. Retail rates, gas infrastructure, permitting, emissions rules and incentives differ by state and country, which means a national or global "market share" figure would blend incomparable projects. Bloom's installed base clusters where the spread is widest: California and the Northeast, where retail rates are high and outages costly, and South Korea, where SK ecoplant is the partner and the country hosts the largest concentration of Bloom installations outside the United States.
The competitive set is fuel cells plus the machinery they displace
| Player | Technology | Where it meets Bloom |
|---|---|---|
| Bloom Energy | solid oxide fuel cell | US, South Korea and Japan on-site power |
| Doosan Fuel Cell | phosphoric acid, solid oxide under development | Korean site and utility projects |
| FuelCell Energy | molten carbonate | US site and utility-scale projects |
| Plug Power | PEM fuel cell, hydrogen | Hydrogen-forward sites, backup power |
| Caterpillar, Cummins | Gas and diesel reciprocating engines | The default on-site generator |
| GE Vernova, Siemens Energy | Gas turbines | Larger industrial and utility plants |
The fuel-cell layer is concentrated: only a handful of companies worldwide ship solid oxide or comparable platforms at commercial scale, and fewer still have a multi-year operating record a customer's insurer will accept. Barriers at that layer are high. The stack is a materials-science problem refined over more than a decade of field data, the service network is expensive to replicate, and each new jurisdiction brings its own interconnection rules and certifications. A new entrant must match Bloom's fleet history before a hospital or a hyperscaler will sign a ten-year service contract.
The alternatives layer is the opposite, and it is the layer that matters for pricing. Grid power, gas reciprocating engines and, increasingly, solar paired with storage are abundant, competitively supplied and subject to their own cost curves. Bloom's share of on-site generation as a whole is small, and the barrier that protects it against a Caterpillar engine is not technology but the specific combination of emissions profile, footprint, federal incentives and no-combustion permitting that a fuel cell offers inside a dense urban site. Where land is plentiful and emissions rules are loose, that advantage thins quickly.
For the moat question, this structure cuts two ways. High concentration and hard technical entry barriers at the fuel-cell layer give Bloom a defensible position against other fuel-cell makers, and the service annuity attached to each installation is sticky once the equipment is bolted down. But the customer's alternative is usually not another fuel cell; it is the utility, and the utility's price is set by regulators who have no interest in losing load. Bloom's pricing power therefore depends on a spread that it influences only at the margin, through cost per kilowatt and efficiency gains. The next question is whether the installed base and service backlog turn that spread into returns on capital that survive a period of falling battery and solar costs.
Competitive advantage
Does the company have a moat? Yes, a moderate one, mostly in service contracts
Moat rating: Moderate. Bloom Energy's protection is not the fuel cell itself. Several companies build fuel cells, and for most buyers the practical alternative is the local utility. What Bloom owns is an installed fleet tied to multi-year service contracts, plus more than a decade of field data on how its stacks age in service. That pairing keeps a customer after the equipment is bolted down. It does not set the price before the sale, because the buyer benchmarks Bloom against the retail rate the utility charges for the same kilowatt-hour. A position defended on one side by a short list of credible suppliers and exposed on the other by the grid earns Moderate, not Strong.
The technology layer is concentrated: only a handful of companies ship solid-oxide platforms at commercial scale, and fewer still carry a service record an insurer will accept. Concentration has not yet produced protected returns. Bloom lost money in each of fiscal 2021 through 2025; the annual statements add up to a cumulative net loss of about $0.88 billion and a cumulative operating loss of about $0.50 billion. The first nine months of fiscal 2026 earned $0.48 billion of operating income. Three quarters test a moat; they do not yet establish one.
Market position: leader of a narrow layer, price taker in the wider market
Market share and ranking: Moderate. Bloom competes in on-site solid-oxide generation for facilities that need firm power around the clock. At that layer it is a commercial-scale supplier with an installed base in the United States and South Korea, and all three of its revenue lines run through the same fleet: product and installation, service, and electricity sold under long-term contracts. The layer is narrow. Most on-site generation is still reciprocating engines, gas turbines or the grid, and Bloom's revenue is small against those alternatives. Leading a segment while taking price from a market you do not lead is a moderate position.
Stability: Moderate. Revenue rose in every fiscal year from 2021 through 2025, from $0.97 billion to $2.02 billion, and the first nine months of fiscal 2026 brought $3.05 billion against $1.25 billion in the same stretch a year earlier, a 144% increase computed from the quarterly statements. A share loser can post that if the pool grows faster, so revenue alone does not prove share gains. What it does show is five years of uninterrupted growth in a business with long sales cycles and no year of contraction, which reads as a stable-to-improving position. Moderate, not Strong, because stability here is inferred from revenue and the installed base instead of measured share.
Profitability: five loss years, then three strong quarters
| Period | Revenue, USD bn | Gross margin, % | Operating income, USD bn | Net income, USD bn | Free cash flow, USD bn |
|---|---|---|---|---|---|
| FY2021 | 0.97 | 20.1 | -0.12 | -0.16 | -0.11 |
| FY2022 | 1.20 | 12.4 | -0.26 | -0.30 | -0.31 |
| FY2023 | 1.33 | 14.8 | -0.21 | -0.30 | -0.46 |
| FY2024 | 1.47 | 27.5 | 0.02 | -0.03 | 0.03 |
| FY2025 | 2.02 | 29.0 | 0.07 | -0.09 | 0.06 |
| 9M FY2026 | 3.05 | 33.1 | 0.48 | 0.52 | 0.04 |
Return on capital: Moderate. The fiscal 2021-2025 record falls below the 6%-8% after-tax return on invested capital that usually signals no advantage at all. Estimated from the statements: fiscal 2025 operating income of $0.07 billion, with no tax provision, over invested capital of about $1.01 billion (equity of $0.77 billion plus $2.69 billion of debt less $2.45 billion of cash) comes to roughly 7%. Fiscal 2022 and 2023 were deeply negative. Equity is not a useful check on the early years, when it hovered near zero or below. The nine months to September 30, 2026 change the picture: $0.48 billion of operating income on invested capital of about $1.98 billion is a 24% return for nine months, not annualized. Leverage is not doing the work; cash of $2.49 billion nearly matches debt of $2.54 billion.
Margins: Moderate. Gross margin rose from 12.4% in fiscal 2022 to 29.0% in fiscal 2025 and 33.1% in the first nine months of fiscal 2026; net margin went from -25.1% to -4.4% to 17.0% across the same three points. Scale spreading fixed costs explains part of that, and price may explain another part. The accounts do not separate the two, which is exactly what a moat judgment needs. Against engine and turbine makers Bloom is small, so its scale advantage exists inside the fuel-cell layer and nowhere else. Cash conversion is the loose thread: nine-month operating cash flow of $0.27 billion against net income of $0.52 billion, with inventory up from $0.64 billion at the start of the year to $0.93 billion at September 30.
Overall rating: Moderate, capped by a short profit record
Overall rating: Moderate. The barrier is real at the technology layer: a short supplier list, a service record that takes years to build, and an installed base that is expensive to rip out. What is not yet established is whether that barrier shelters profit through a full cycle. Fiscal 2021-2025 produced negative operating income in four of five years, and the turnaround rests on nine months of results and on a gross margin that could reflect manufacturing cost declines as easily as pricing power. The two sub-ratings that carry the most weight, return on capital and margins, are Moderate for the same reason: the direction is right and the duration is short. That is why the overall rating sits at Moderate instead of the Fairly Strong that a stable, growing installed base can otherwise support.
Where the advantage comes from and how long it can last: a sticky annuity behind a thin price spread
Bloom's edge rests on three things: what it knows about making stacks, what it costs a customer to leave, and how cheaply it can service a dense fleet. They are worth separating, because only the second is hard to copy. The test that matters is the price Bloom gets per kilowatt-hour over the life of an installation against what the local utility charges. That spread has to be inferred from the consolidated gross margin, which mixes product price, manufacturing cost and service economics, so treat it as a proxy that cannot fully separate price from cost.
Supply side: a falling cost curve that rivals can follow
Supply-side advantage: Moderate. Bloom's manufacturing edge is process know-how: seals, interconnects, coatings and thermal cycling, refined with field data from a fleet that has run for years. A rival can hire the engineers but cannot buy the operating hours. The cost curve shows up in the margin. Gross margin went from 12.4% in fiscal 2022 to 33.1% in the first nine months of fiscal 2026, and roughly $125 million of the latest nine-month gross profit came from the higher rate rather than from volume (an estimate: $3.05 billion of revenue at fiscal 2025's 29.0% rate gives $0.885 billion, against $1.01 billion reported). That is what the advantage was worth in nine months.
It is not clearly a barrier. Learning transfers with people, and a stack is only part of what a buyer pays for: installation, gas interconnection, permitting and service make up much of the installed cost and are all local. What would break it is a competitor shipping a warranted stack at a comparable cost per kilowatt, which would pull the consolidated margin back toward the 16% average of fiscal 2021-2023.
Demand side: captivity begins after the equipment is installed
Demand-side advantage: Moderate. Customers stay because leaving is expensive. The contract does real work: a multi-year service agreement, proprietary parts, remote monitoring and warranty terms that make third-party servicing impractical. The switch itself is a construction project that interrupts firm power to a hospital, a fab or a data hall, and it goes back through permitting. Captivity of that kind is genuine, and it shows up as the service line repeating every year.
The limitation is timing. Captivity starts at commissioning, and the decision that matters happens earlier, when the buyer compares Bloom with an engine, a turbine or the utility. At that moment Bloom is one bidder among several with little room to raise price, because the benchmark is the retail rate. There is no network effect to lean on: one customer's modules do not make another's more valuable, so growth brings scale without self-reinforcement. Partial captivity earns Moderate.
The adverse fact to watch is stack replacement. When the early cohorts reach mid-life refurbishment, service cost spikes on equipment already priced under contract. If consolidated gross margin holds through that, the annuity is real. If margin sags only when replacement work peaks, the ten-year contract protected a warranty obligation, not a profit stream.
Scale: service density is the barrier, not factory size
Economies of scale: Moderate. The fixed costs are depots, field technicians, applications engineers and certification in each utility territory. Those spread over the fleet within a region, and the arithmetic is unforgiving for a small rival: double the modules a technician can reach within an hour's drive and the labor cost per module served falls by half. Bloom's fleet clusters in California, the Northeast and South Korea, where that density exists.
The barrier is modest for two reasons. A new entrant must fund depots before it has the fleet, which is the classic scale trap, but the entrant here would likely be an engine maker or a Korean industrial group with deep pockets, and Bloom's Korean presence runs through a local partner, so part of its geographic reach is not an asset it owns outright. The leverage also cuts both ways: if installations slow, the same fixed service base spreads over fewer modules while contract prices stay put. What would break the scale advantage is a stalled order book, which turns operating leverage into margin compression without any rival doing anything.
Durability: the grid and batteries set the ceiling
The substitute that matters is not another fuel cell; it is the utility and, increasingly, solar paired with batteries. Where land is cheap and the grid is dependable, storage wins on cost and Bloom's spread disappears. Where the site is land-constrained and emissions-constrained, the fuel cell's small footprint and low local emissions are worth more than a few cents per kilowatt-hour. That niche is defined by retail rates and by regulation, neither of which Bloom controls. The federal investment tax credit improves the customer's payback but is available to competing technologies on similar terms, so it lifts demand without protecting price.
So the moat is real and conditional. It protects the service tail of each sale and the small circle of credible suppliers, and it leaves the next sale exposed to whatever the grid charges. The evidence that would settle it is gross margin through a stack-replacement cycle, alongside the share of revenue that repeats as service rather than arriving with new equipment. If both hold as volume grows, the annuity is genuine. If margin holds only while new units ship, the advantage sits in the factory, where it is exposed to any rival that matches Bloom's cost per kilowatt.
Generated by DeepVal AI. Data and models may contain errors or be outdated. For research only; not personalized investment advice.
Continue reading the full report
- Comparison and summaryHow Bloom Energy Corp compares with its closest rivals, and our moat rating
Free to sign up · 20 research points · unlocking this report uses 10
20 points cover 2 full reports. No card needed.
Research reflects its generation date. Preview the first two modules.
More on Bloom Energy Corp
Bloom Energy Corp
Business, moat and key questions
Open company page