
Bloom Energy
Hardware sales of Energy Servers (bulk of revenue) plus recurring service/maintenance contracts, electricity (managed-services/PPA) revenue, and installation; large-scale project financing increasingly carried by partners (e.g. Brookfield) rather than Bloom's own balance sheet.
- 2026-08-04This market capitalisation previously read ~$93B (as of June 24, 2026). Restated to $64.30B on this refresh, roughly 31% lower.
- 2026-08-04This share price previously read ~$342 (as of June 25, 2026). Restated to $218.32 on this refresh, roughly 36% lower.
Sources — 16 figures with citations
- Q2 FY2026 revenue and growthfiled2026-06-30$1,065.4M vs $401.2M (+165.5%); product revenue $935.4M (+215.4%); H1 $1,816.4M vs $727.3M (+149.8%)sec.gov — Condensed Consolidated Statements of Operations, 10-Q filed 2026-07-28
- Q2 FY2026 marginsfiled2026-06-30Gross profit $355.6M = 33.4% GAAP (34.3% non-GAAP); operating income $182.2M = 17.1% GAAP (22.5% non-GAAP); net income to common $196.3M; diluted EPS $0.62 GAAP / $0.78 non-GAAP; adjusted EBITDA $253.4Msec.gov — Q2'26 earnings release, Exhibit 99.1 to the 2026-07-28 8-K; ties to the 10-Q income statement (355,572/1,065,365 = 33.4%; 182,237/1,065,365 = 17.1%)
- Customer concentration (corrected disclosure)filed2026-06-30Q2 FY2026: one customer, not a related party, was ~73% of total revenue. H1 FY2026: two customers were ~44% and ~21%, the second being a related partysec.gov — 10-Q/A filed 2026-07-29 — the amendment exists solely to correct a three-months/six-months transposition in this Note 1 concentration-of-risk paragraph
- Related-party revenue splitfiled2026-06-30$2.8M in Q2 FY2026 vs $376.1M for H1 FY2026 — almost all related-party revenue landed in Q1sec.gov — Footnote 1 to the Condensed Consolidated Statements of Operations
- Cash, debt and net positionfiled2026-06-30Cash & equivalents $2,666.9M + restricted cash $21.6M ($1.05M current + $20.6M non-current) = $2,688.5M; recourse debt $2,475.4M ($4.7M current + $2,470.7M non-current); non-recourse debt $2.6M; financing obligations $206.5Msec.gov — Condensed Consolidated Balance Sheets. Net cash = 2,688.5 - 2,478.0 = +$210.5M; including financing obligations = +$4.0M (derived arithmetic)
- H1 FY2026 cash flow and capexfiled2026-06-30Operating cash flow $300.0M (vs -$323.8M in H1 FY2025); purchases of property, plant & equipment $77.8M (vs $21.5M); investments in unconsolidated affiliates $22.8Msec.gov — Condensed Consolidated Statements of Cash Flows
- Capex intensity and FCF margin (derived)derived2026-06-30H1 capex intensity 4.3%; H1 FCF $222.2M = 12.2% FCF marginsec.gov — 77.8/1,816.4 = 4.28%; 300.0 - 77.8 = 222.2; 222.2/1,816.4 = 12.2%. H1 used rather than Q2 because the cash flow statement is presented on a six-month basis only
- FY2025 revenue base and TTM (derived)derived2025-12-31FY2025 GAAP revenue $2,024.0M; TTM revenue $3,113.1Msec.gov — FY2025 revenue $2,024.0M is stated in the Q2'26 supplemental deck's five-year GAAP-to-non-GAAP reconciliation (and matches the SEC XBRL 'Revenues' fact for FY2025). TTM = 2,024.0 + 1,816.4 - 727.3 = 3,113.1
- FY2026 guidance (raised)filed2026-07-28Revenue $3.9-4.2B (~+100% YoY at midpoint); non-GAAP gross margin ~34%; non-GAAP operating income $800-900M; non-GAAP EPS $2.55-2.85sec.gov — 'Guidance' section of the Q2'26 earnings release; the supplemental deck repeats it with FY2025 comparatives ($2.0B revenue, 30% non-GAAP GM, $221M non-GAAP operating income)
- Stock-based compensation and dilutionfiled2026-07-22SBC $56.4M in Q2 / $100.4M in H1; weighted-average diluted shares 323.3M vs 287.3M basic in Q2; shares outstanding 294,527,346 as of 22-Jul-2026 (280,045,459 at 31-Dec-2025); $144.6M of Green Notes converted in H1sec.gov — Share count from the 10-Q/A cover page; SBC and conversions from the cash flow statement and the supplemental reconciliation
- Balance-sheet order-book proxiesfiled2026-06-30Deferred revenue & customer deposits $445.0M (from $143.8M at 31-Dec-2025); contract assets $428.3M (from $241.2M); customer consideration asset $306.5M newly recognised; accumulated deficit $3,721.0Msec.gov — Condensed Consolidated Balance Sheets; the customer consideration asset is described in the filing as related-party upfront share-based consideration payable to a customer's customer
- Share price (close)market2026-08-03$218.32 close on 2026-08-03, +$12.51 (+6.08%) on the daystockanalysis.com — Closing print, 4:00 PM EDT 2026-08-03. Not an intraday high
- Market cap and implied multiplesmarket2026-08-03$64.30B market cap on 294.53M shares; ~20.7x TTM revenue; ~15.9x the FY2026 guidance midpoint of $4.05Bstockanalysis.com — Market cap from the provider at the 2026-08-03 close; multiples derived: 64.30/3.113 = 20.7x, 64.30/4.05 = 15.9x
- Brookfield financing frameworkfiled2026-06-30Partnership expanded to a $25B framework on 30-Jun-2026, a fivefold increase from the $5B framework announced Oct-2025; sits within Brookfield's AI Infrastructure Fund (launched Nov-2025, $100B target)investor.bloomenergy.com — Bloom Energy's own investor-relations press release. It is a financing framework for project capital, NOT booked revenue or backlog
- Short-seller allegation and company responsemarket2026-07-27Hunterbrook Capital, disclosing a short position, alleged Bloom downplayed reliance on Chinese scandium-oxide suppliers and that scaling ~1GW (2026) toward 5GW/yr would need ~220 metric tons of scandium oxide against ~240 tons of global annual production. Bloom called the accounting and financial-metric claims 'false and misleading'hntrbrk.com — Hunterbrook's own published investigation is the primary source for the allegation; the company rebuttal is reported in the coverage at https://finance.yahoo.com/markets/stocks/articles/short-seller-hunterbrook-attacked-bloom-133002104.html. Neither side's supply-chain claim is independently verified here
- Drawdown contextmarket2026-07-28All-time high $351.28 on 25-Jun-2026; closed $184.89 on 24-Jul-2026 (-14.9% that day); traded down ~13% to ~$163.04 intraday on 28-Jul-2026 ahead of the print; ~43% one-month decline into earnings247wallst.com — The $163.04 figure is explicitly intraday and is NOT used as the snapshot price
The thesis on this name
State of Data-Center Power
Solid-oxide fuel cells have crossed from niche backup to primary on-site generation for power-starved datacenters because they deploy in ~90 days vs years for grid interconnect — the speed-to-power arbitrage against the ~2,300GW queue. Bloom signed $7.65B of datacenter-related contracts in a ~90-day window in early 2026, anchored by a 2.8GW Oracle deal, a 1GW AEP agreement, and a $5B Brookfield deployment partnership (fact). The asymmetry: the market still prices BE as a speculative clean-tech rather than the bankable speed-to-power solution hyperscalers are now signing in GW blocks. High-potential and higher-risk (gas-fed economics, execution scale-up) — a satellite, not a core toll.
State of Data-Center Power
Real on-site fuel-cell demand (+219% YTD, $7.65B of datacenter deals) but the stock has already priced the win — momentum, not margin of safety.
Earnings, margins, COGS & capex
Bloom inflected hard in Q1 FY26: revenue $751M (+130% YoY) driven by product revenue +208% as AI-datacenter on-site power orders converted, swinging GAAP net income to +$70.7M from a -$23.8M loss a year earlier (fact). Non-GAAP gross margin reached 31.5% and management raised FY26 guidance to $3.4–3.8B revenue with ~34% non-GAAP gross margin and $1.85–2.25 non-GAAP EPS (guidance). The model is capital-light at the box level (capex ~3.5% of revenue) because project capital is increasingly externalized to partners like Brookfield; the swing factor is manufacturing-ramp execution toward 2 GW capacity and sustaining margin as volume scales.
Income statement — where each revenue dollar goes
% of revenueOf every $1 of revenue, ~67¢ is cost of goods and ~16¢ operating expense, leaving ~17¢ of operating profit (~9¢ net).
Revenue trend
Margins
up (+2.8pp YoY; FY26 guide ~34%)
up sharply (op income $129.7M vs $13.2M YoY)
up — first sustained GAAP profitability (vs loss)
up — OCF +$184M YoY swing to positive
COGS structure
COGS is dominated by the fuel-cell stack bill-of-materials — ceramic electrolyte/anode materials, balance-of-plant (inverters, hot-box assembly), and contract-manufactured components — plus factory labor/overhead at Fremont CA and Newark DE. Margin is gated by volume-driven fixed-cost absorption, materials/tariff exposure (Bloom flagged tariff hits), and warranty/field-service costs on the installed fleet; service COGS has been the historical drag but service gross margin jumped to 18% non-GAAP in Q1 FY26 as fleet reliability improved.
Capex
Low absolute capex (~$26M in Q1 FY26, ~3.5% of revenue) funding manufacturing-capacity expansion toward 2 GW annual run-rate by end-2026 (automation, Fremont/Newark lines). Capital intensity stays modest because customer-side project capital (the power plants themselves) is increasingly financed by partners (Brookfield $5B vehicle) rather than on Bloom's balance sheet.
Latest earnings
Large beat and a guidance raise: revenue of $1,065.4M came in roughly 29% above analyst estimates per contemporaneous coverage, and it was the first quarter above $1B. Every headline metric inflected — first meaningful GAAP operating income ($182.2M vs a $3.5M loss) and first positive H1 operating cash flow ($300.0M vs -$323.8M). The stock had already fallen ~43% in the month into the print and slipped further on the day despite the beat, then recovered to $218.32 by 3-Aug
FY2026 guidance RAISED on 28-Jul-2026: revenue $3.9-4.2B (~+100% YoY at midpoint, from $2,024.0M in FY2025); non-GAAP gross margin ~34% (~+374bp); non-GAAP operating income $800-900M (~3.8x FY2025's $221.0M); non-GAAP EPS $2.55-2.85 (~3.6x). Implies an H2 run-rate of roughly $2.1-2.4B against $1.82B delivered in H1
- Product revenue
- $935.4M, +215.4% YoY (88% of total revenue)
- Single-customer concentration, Q2
- One customer, NOT a related party, was ~73% of total revenue
- Customer concentration, H1
- Two customers were ~44% and ~21% of revenue; the 21% one is a related party
- Related-party revenue
- $2.8M in Q2 but $376.1M in H1 — i.e. essentially all of it landed in Q1
- Adjusted EBITDA
- $253.4M vs $41.2M a year ago (~6.1x)
- Stock-based compensation
- $56.4M in Q2 (5.3% of revenue); $100.4M in H1
- Deferred revenue & customer deposits
- $445.0M total (current $327.1M + non-current $117.9M), up from $143.8M at 31-Dec-2025
- Contract assets
- $428.3M (current $365.5M + non-current $62.8M), up from $241.2M at 31-Dec-2025
- Customer consideration asset
- $306.5M newly recognised (current $91.0M + non-current $215.5M) — related-party upfront share-based consideration payable to a customer's customer
- Accumulated deficit
- $3,721.0M, still larger than the current market value of equity book ($1,641M total stockholders' equity)
- Diluted share count
- 323.3M weighted-average diluted in Q2 vs 287.3M basic — ~13% dilution overhang from convertibles
Growth drivers
- AI-datacenter on-site power — $7.65B of fuel-cell datacenter deals signed in the sector and a backlog approaching ~$20B (~6x revenue run-rate); demand +219% YTD per board framing
- Grid-interconnection bottleneck — utilities quote multi-year hookup queues, so hyperscalers buy Bloom for 'power in ~90 days' (Oracle/OCI deal)
- Capacity doubling to 2 GW by end-2026 unlocks unit volume and fixed-cost absorption → the engine of the margin guide to ~34%
- Marquee anchor deals — AEP 1 GW order, Brookfield $5B financing partnership, Equinix 100MW+ across 19+ datacenters, SK ecoplant international distribution
- Service-margin inflection — recurring service contracts turning profitable (18% non-GAAP gross margin) as installed base and fleet reliability grow
- Behind-the-meter / electricity (managed-services) recurring revenue and replacement-stack cycle on a growing fleet
Reported financials — SEC EDGAR
Audited GAAP figures pulled from SEC filings · latest filing 2026-02-09. The audited primary-source spine — not financial advice.
Revenue — annual (GAAP)
Margins & balance sheet — FY’25
Bull & bear
Bloom crossed from a story into a genuinely profitable business in one quarter: first $1B+ revenue quarter, first real GAAP operating income ($182.2M at 17.1%), first positive H1 operating cash flow ($300.0M), and a guidance raise to ~100% FY2026 growth. Behind-the-meter fuel cells are being bought as the fastest path around multi-year utility interconnection queues, and Brookfield's $25B financing framework de-risks the customer's capital.
- The inflection is in the filings, not the pitch: GAAP operating income of $182.2M against a $3.5M loss a year ago, with gross margin +668bp to 33.4% and non-GAAP operating margin +15.4pt to 22.5%. Operating leverage is arriving exactly where the model said it would.
- Cash generation flipped: Q2 operating cash flow $226.4M, H1 $300.0M versus a $323.8M H1 outflow last year — a $624M swing. With H1 capex of only $77.8M, this is a capital-light manufacturer relative to the power it sells.
- Guidance was raised, not merely reaffirmed: FY2026 revenue to $3.9-4.2B and non-GAAP operating income to $800-900M — roughly 3.8x FY2025's $221.0M. Management raised into a stock that had already fallen 43% in a month.
- Demand-side validation is broad on management's account: all major US hyperscalers plus over a dozen neoclouds, AI labs and colocation operators have validated and approved Bloom's power solutions.
- The Brookfield partnership was expanded fivefold to a $25B framework on 30-Jun-2026 (from $5B in Oct-2025), sitting inside Brookfield's $100B-target AI Infrastructure Fund — meaning the customer's financing constraint, historically Bloom's biggest sales obstacle, is being solved by a third party's balance sheet.
- Balance sheet is no longer the risk it was: $2,688.5M of cash and restricted cash against $2,478.0M of debt is a small net cash position, and forward order visibility is building — deferred revenue and customer deposits tripled to $445.0M and contract assets rose to $428.3M since year-end.
The blowout quarter rests on one customer: a single non-related-party customer was ~73% of Q2 revenue, and two customers were 44% and 21% of H1. That is project revenue, not a run-rate, and it makes the raised H2 guide a concentration bet. Layered on top are two short-seller attacks alleging a hidden Chinese scandium-oxide dependency, a regulatory rejection of a pipeline tied to the largest disclosed growth contract, and a 13% convertible dilution overhang.
- Single-customer concentration is extreme and newly disclosed: ~73% of Q2 revenue from one customer, and 44% + 21% from two customers in H1. Bloom filed a 10-Q/A the day after the 10-Q specifically to fix a three-months/six-months transposition in exactly this disclosure — the concentration is the most load-bearing and least-emphasised fact in the release.
- Revenue is lumpy project revenue, so the raised FY2026 guide ($3.9-4.2B) implies an H2 of roughly $2.1-2.4B against $1.82B in H1 — a further step-up that depends on specific site deliveries landing in the period, not on a subscription base.
- The economics are still concentrated in one line: product is 88% of revenue at 36.5% gross margin, while installation runs at a NEGATIVE 3.6% gross margin and electricity revenue actually shrank YoY ($10.0M vs $12.8M). The recurring/service tail is small relative to the hardware pulse.
- Hunterbrook Capital, disclosing a short position, alleged Bloom downplayed reliance on Chinese scandium-oxide suppliers and challenged the physical feasibility of scaling installed capacity from ~1GW in 2026 toward 5GW/year — its arithmetic put that need at ~220 metric tons of scandium oxide against ~240 tons of global annual production. Bloom categorically rejected the accounting and financial claims as 'false and misleading'; the supply-chain question is unresolved either way.
- A second regulatory rejection of a natural-gas pipeline tied to Bloom's largest disclosed growth contract removed a physical prerequisite for that revenue. Fuel supply and permitting, not fuel cells, are the binding constraint on grid-bypass power.
- Dilution is material and ongoing: 323.3M weighted-average diluted shares versus 287.3M basic (~13%), share count up from 280.0M at year-end to 294.5M by 22-Jul-2026, and $144.6M of Green Notes converted in H1. Non-GAAP EPS is also flattered by $56.4M of quarterly SBC.
- The $306.5M 'customer consideration asset' — related-party upfront share-based consideration payable to a customer's customer — is an unusual structure to book as an asset, and it means part of the demand was purchased with equity-linked consideration rather than won purely on price and performance.
- Valuation still requires the ramp: ~$64.3B market cap on $3.11B TTM revenue is ~20.7x sales (~15.9x the FY2026 guide midpoint) for a hardware business with a $3.72B accumulated deficit and a customer list of one that matters this quarter.
What it is worth
Multiple-based sanity check anchored to FY26 guidance + analyst targets (forward P/E and EV/Sales), cross-checked against sell-side range. Not a full DCF — backlog visibility is high but out-year margin and the AI-capex cycle are the swing variables.
~$150–200
multiple compresses toward ~20–25x EPS on any AI-capex cooling, deal slippage, margin miss, or competitive substitution
~$237 (≈24-analyst average target)
strong growth but multiple normalizes toward a premium-but-sane level as AI-capex matures
~$335+ (RBC Buy $335)
if capacity ramps, margins hit ~34%+, and the AI-power TAM keeps expanding the multiple holds/expands
At ~$342 the stock trades ~140x FY26 non-GAAP EPS midpoint (~$2.05) and ~27x FY26 sales — pricing flawless multi-year execution; the average analyst target (~$237) sits below the price, so the win is in the stock (momentum, not margin of safety).
SWOT
Strengths
- Technology lead in commercial SOFC — fuel-flexible (gas today, hydrogen-ready), high efficiency, fast-deployable — winning while PEM/hydrogen peers (Plug, Cummins) cut investment
- Demand inflection now hitting the P&L — +130% revenue, GAAP profitability, +$184M operating-cash-flow swing in a single quarter
- Marquee, balance-sheet-light backlog — AEP 1 GW, Brookfield $5B financing, Oracle, Equinix 100MW+ — ~$20B backlog (~6x run-rate)
- Speed-to-power moat — ~90-day deployment vs multi-year utility interconnection is a structural buyer pull, not a price pitch
- Margin trajectory improving (service gross margin 13.3%→ positive; guide to ~34% non-GAAP gross margin)
Weaknesses
- Customer concentration in a single hot end-market (AI datacenters) and a handful of mega-customers/partners
- Tariff and materials cost exposure flagged by management; COGS sensitive to input prices and volume absorption
- Net-debt balance sheet (cash ~$2.5B vs ~$2.6–3.0B debt) despite the cash-flow turn; recourse debt $2.60B
- History of GAAP losses, dilution, and lumpy quarterly revenue recognition tied to large project timing
- Reliance on natural gas means the 'clean' label is contested (CO2 emissions vs zero-carbon alternatives)
Opportunities
- AI datacenter power deficit could run for years — Bloom is positioned as default behind-the-meter primary power, not just backup
- Hydrogen/electrolyzer optionality if green-H2 economics arrive (same SOEC platform)
- International expansion via SK ecoplant (Korea) and new geographies; data-center power is global
- Capacity build to 2 GW (and beyond) plus operating leverage could compound EPS faster than revenue
- Recurring service + electricity revenue and replacement-stack cycle on a fast-growing installed base
Threats
- Competing power solutions for datacenters — natural-gas turbines (GE Vernova, Caterpillar/Cummins gensets), grid upgrades, on-site solar+storage, and nuclear/SMRs
- Valuation risk — ~140x forward EPS / ~27x sales prices in years of flawless execution — any AI-capex cooling compresses the multiple fast
- Policy/tax-credit dependency (ITC/IRA-style credits) — adverse changes raise customer cost of ownership
- Execution/ramp risk doubling capacity to 2 GW; supply-chain and quality scaling can dent margins
- Hyperscalers could vertically integrate or standardize on a rival power architecture, eroding Bloom's incumbency
Moats, dependencies & bottlenecks
Moats
durable while interconnection queues persist (multi-year), erodes if grid/turbine supply catches up ~90-day on-site deployment vs multi-year utility hookup is the core buyer pull (Oracle/OCI deal); a workflow + relationship advantage more than a hard tech lock-in
real know-how (electrolyte chemistry, hot-box, fleet reliability) but peers and gas turbines compete Bloom is the leading commercial SOFC player while PEM/hydrogen rivals retreat; lead is in efficiency, fuel-flexibility, and a 2 GW manufacturing base
recurring service/replacement-stack revenue grows with fleet, raises switching cost once a site standardizes on Bloom Service gross margin turned positive (18% non-GAAP); the longer Bloom is on-site the stickier it gets
relationship/contract-based, not exclusive Brookfield $5B, AEP 1 GW, Equinix, SK ecoplant give scale capital and distribution rivals lack today, but these are negotiable, not structural barriers
sentiment-driven, reverses if a deal slips or a rival lands a marquee hyperscaler Mindshare premium is reflexive with the stock and not defensible on its own
Dependencies
Revenue is now levered to a single demand theme; an AI-capex pause or grid catch-up directly cuts orders and the multiple
AEP, Equinix, Brookfield, SK ecoplant) Backlog concentrated in a handful of names/partners; loss or slippage of one materially moves the model
Customer TCO and Bloom's value prop depend on gas availability/price at site; hydrogen optionality is not yet economic
Customer payback economics lean on credits; adverse policy or tariff changes raise cost of ownership and dent demand
Ceramic materials, inverters/balance-of-plant, and contract manufacturing must scale 2x without quality/margin slippage
Advantages
- First-mover incumbency as the proven, bankable SOFC vendor for datacenters while hydrogen/PEM rivals retreat
- Speed: ~90-day deployment beats utility interconnection by years — a differentiator buyers will pay up for
- Fuel flexibility (gas now, hydrogen-ready) hedges the energy-transition path and widens the addressable market
- Capital-light model: partners fund the power plants, Bloom keeps the high-margin equipment + service annuity
- Demonstrated financial inflection (GAAP profit, positive OCF) gives credibility and self-funding capacity rivals lack
- Marquee reference customers (Oracle, Equinix, AEP) create a credential flywheel for the next datacenter buyer
Weaknesses
- Extreme valuation (~140x fwd EPS, ~27x sales) leaves no margin of safety and trades above average analyst target
- Concentration on one end-market (AI datacenters) and a few mega-customers/partners
- Net-debt balance sheet and a history of GAAP losses, dilution, and lumpy revenue recognition
- Natural-gas dependence undercuts the zero-carbon narrative and exposes the thesis to decarbonization-policy shifts
- Tariff/materials cost exposure flagged by management can pressure the margin-expansion story
- Reflexive 'AI power' sentiment premium can reverse violently on any single deal slip or capex headline
Bottlenecks
- Manufacturing capacity — must roughly double to 2 GW by end-2026 to fulfill the ~$20B backlog; throughput is the binding constraint on revenue conversion
- Component/materials supply (ceramic electrolyte, inverters/balance-of-plant) and tariff-exposed inputs limit how fast COGS scales down
- Skilled field-deployment and interconnection/permitting labor to install at hyperscale pace across many sites
- On-site gas availability and local permitting/air-quality approvals can gate individual datacenter deployments
- Project capital — mitigated by Brookfield, but very large multi-GW orders still require continued third-party financing capacity
Top signals & trends
Top signals
~$20B backlog and $7.65B datacenter deals are the engine; watch for new hyperscaler wins vs any push-out or cancellation as the leading tell
Margin is the bridge from revenue growth to the EPS that justifies the multiple; a miss here de-rates the stock fast
Throughput is the binding constraint; ramp slippage or quality/warranty issues would cap revenue and dent margins
Stock trades above consensus fair value; the win is priced in — momentum, not margin of safety
Bloom is now a high-beta AI-capex proxy; any sign of capex normalization compresses the multiple
Customer economics and Bloom's COGS both hinge on credits and tariffs; watch IRA/ITC and trade-policy moves
Trends
The core tailwind — datacenters need on-site power now, which is exactly Bloom's pitch; durable for the medium term
Primary on-site power (not just backup) normalizes SOFCs as a standard datacenter asset class (AEP, Brookfield)
Validates Bloom's gas-compatible bet and removes competitors near-term, but weakens the long-run zero-carbon story
GE Vernova turbines and future SMRs target the same demand; substitution risk as supply of alternatives scales
Tax credits help demand, but tightening carbon rules pressure the natural-gas value prop over time
Powers the rally now; a high-beta liability if AI-capex sentiment turns
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
Primary fuel for the installed fleet; on-site gas availability gates deployments
Ceramic / electrolyte & balance-of-plant component vendors Stack materials, inverters, power electronics for the hot-box — key COGS and tariff-exposed inputs
Contract manufacturers & automation suppliers Support the Fremont CA + Newark DE lines scaling to 2 GW capacity
Korea partner — also a manufacturing/distribution supplier under the long-running JV
Datacenter on-site power with ~90-day deployment commitment to accelerate AI infrastructure
100MW+ across 19+ IBX datacenters in six states — flagship recurring datacenter customer
1 GW order — utility deploying Bloom for grid-adjacent/datacenter load
The demand driver behind the $7.65B of datacenter deals and ~$20B backlog
Gas turbines + grid equipment — the largest substitute for datacenter on-site/primary power; deep utility relationships and scale
Large-engine gensets and backup/prime power for datacenters; entrenched distribution and service network
Power-generation gensets and (scaled-back) hydrogen/fuel-cell efforts; competes in distributed gas power
Molten-carbonate/SOFC carbonate platform; ~4 GW pipeline ~90% datacenter proposals but far smaller scale and weaker financials
PEM/hydrogen ecosystem; retreating from the space Bloom dominates, weak economics — competitor more in narrative than datacenters
Grid-scale + nuclear (incl. SMR) datacenter power deals; a longer-dated substitute for behind-the-meter fuel cells