--- title: "BE (Trans): 3rd Guidance Hike, Gearing Up for the Post-5 GW Phase" type: "Topics" locale: "en" url: "https://longbridge.com/en/dolphin/post/44506473.md" description: "Q3 revenue was $1.24bn (+138% YoY). Full-year revenue guidance was raised for the third time this year to $4.4–4.6bn (midpoint ~+122% YoY).However, full-year GPM guidance remains ~34%, and operating cash flow saw a $27mn net outflow in the quarter. The company has begun preparing for expansion beyond its 5GW capacity." datetime: "2026-10-22T22:58:10.000Z" locales: - [en](https://longbridge.com/en/dolphin/post/44506473.md) - [zh-CN](https://longbridge.com/zh-CN/dolphin/post/44506473.md) - [zh-HK](https://longbridge.com/zh-HK/dolphin/post/44506473.md) author: "[Dolphin Research](https://longbridge.com/en/dolphin.md)" generator: "portal-rs" --- # BE (Trans): 3rd Guidance Hike, Gearing Up for the Post-5 GW Phase **The following is Dolphin Research's recap of** $Bloom Energy(BE.US) **Q3 2026 earnings call** **I. Key takeaways from the print** 1) Full-year guidance (third raise this year) a. Revenue: raised to $4.4–$4.6 bn, midpoint implies ~+122% YoY. At the Q1 call, management guided to ~+60% growth. b. Profit: non-GAAP OP raised to $960–$1,030 mn, with midpoint OPM at 22% vs. 14% in Feb guidance. Non-GAAP diluted EPS at $3.10–$3.40. c. GPM: non-GAAP GPM maintained at ~34% (2025: 30%, early-year guide ~32%). This already reflects Q4 mix and segment-level margins. d. Basis: guidance is a probability-weighted view of backlog (by customer, site readiness, conversion timing) plus pipeline and in-quarter urgent power needs. Timing of a few year-end deliveries will determine where results land within the range. 2) This quarter's key metrics a. Aggregate: revenue $1.24 bn (+138% YoY, +16% QoQ). OP $286 mn (>+500% YoY), surpassing full-year 2025 in a single quarter; OPM 23.1% vs. 8.9% a year ago. Opex $141 mn, up only 26%. b. Other profit metrics: Adj. EBITDA $299 mn (24% of revenue). Non-GAAP diluted EPS $0.94 vs. $0.15 last year. c. Segments: product revenue $1.07 bn (+179% YoY), first single quarter above $1 bn, accounting for 87%. Services revenue $67.6 mn (+15% YoY). Installation revenue $84.9 mn (+29% YoY, +66% QoQ). d. Margins: product GPM 38.4% (+250 bps YoY, +120 bps QoQ). Services GPM 21.4% (+700 bps YoY), above 20% for two straight quarters. Total GPM 34.5% (+412 bps YoY), partially offset by timing swings in installation mix. 3) Cash flow and capex a. Operating cash flow: net outflow of $27 mn, as net income was offset by higher inventory, contract assets, and AR. Product revenue is recognized upon transfer of control, while invoicing follows contract milestones and cash collection lags revenue recognition. b. Capex: $153 mn in the quarter, $231 mn YTD, already 4x the full year 2025 level. Spending is focused on equipment and capacity within existing Fremont and Newark sites, with incremental investments aligned to demand. c. FCF: -$181 mn in the quarter, +$42 mn YTD. Period-end cash at $2.5 bn, roughly flat vs. the start of the year; the company expects to continue investing working capital to support growth. **II. Earnings call details** **2.1 Management highlights** 1) Industry view: on-site generation is expanding from data centers to the broader economy. Grids were designed for distributed loads and struggle with concentrated DC demand; cost-plus pricing and franchised models under-incentivize innovation, and flat AI power usage over the past two decades led to underinvestment. b. A survey of 15 sub-industries and 200+ commercial and industrial firms shows a quarter of facilities expect fully on-site power by 2030, rising to 35% by 2035. IEA projects global electricity demand to grow 40% by 2035. c. Management expects the power market to double from roughly $3 tn today to about $6 tn in the next decade. Countries with insufficient grid infrastructure will leapfrog to distributed on-site power. 2) Deliveries and demand: backlog is growing faster than revenue, with broader customer, industry, and geographic mix. The first Oracle project in Utah began delivering power in under two months 15 months ago; the Nebius project in New Jersey secured planning approval within three months of signing, and the first block will be customer-ready two months post-approval. c. Virginia Gov. Abigail Spanberger named non-combustion fuel cells in the 2026 energy plan, citing far lower local emissions vs. gas-fired generation. Virginia is the most concentrated data center hub globally. 3) Five commercial advances this quarter (expanding TAM): a. Military: the U.S. Army conditionally awarded partner Indelible a long-term lease to build and operate generation at Army bases using Bloom servers; critical civil-military infrastructure (bases, water plants, hospitals, airports) is viewed as a growth area. b. CCS: AEP disclosed it will use 700 MW of Bloom servers with carbon capture and storage to provide near net-zero power to Google. Management views CCS as a potential trillion-dollar market. c. Regions: continued expansion in the Rockies, Midwest, and Great Lakes; competitive even in low-power-price Texas, with the first batch shipped this quarter. d. Utilities: in-front-of-the-meter business is seen as a channel, with major orders expected. e. Intl: multiple markets show strong interest, and the company is engaging local partners, with orders likely to be announced soon. 4) On uncertainty: management acknowledged macro instability and market volatility, but reiterated that power demand will keep rising. The company aims to grow through self-funded operations. **2.2 Q&A** **Q: Why expand the Fremont site now? Existing capacity supports 5 GW+; does expansion imply 10 GW+?** A: **No capacity figure or timeline post-expansion; expansion prepares for beyond 5 GW.** Management expects power demand to rise sharply from AI and needs around resilience and security, with similar trends in Asia and Europe. 5 GW is only a milestone on a long-term path and is clearly in sight. Capacity will be ramped gradually with demand, not ahead of it, and will not be a bottleneck for customers securing power. **Q: Guidance raised for revenue but GPM unchanged; beyond mix, what else? Will margins expand further in 2027 as scale grows?** A: **No 2027 guide; current GPM includes start-up costs for expansion, with scale benefits to release over time.** Total GPM rose from 28.7% in 2024 and 30.3% in 2025 to 34.5% this quarter; product pricing remains strong, and cost-down is already reflected in the P&L. New line ramps and expedited deliveries add short-term costs; multiple engineering, procurement, and manufacturing cost programs need time to flow through to COGS. As projects come online, lines mature, and utilization rises, scale benefits should further materialize. **Q: Op. cash flow outflow of $27 mn this quarter—any one-offs or noise?** A: **The single-quarter outflow is 'not the start of a trend'; assess on a full-year basis, as the business is positioned to be FCF-positive and self-funding.** YTD operating cash flow is positive, while increased headcount and capability investments and working capital advanced for customer deployments covered all capex. Self-funding reduces reliance on external financing and execution risk, while preserving flexibility to invest per customer needs. **Q: Many IRPs now include fuel cells; how do you view utility-side (in-front-of-the-meter) opportunities and TAM?** A: **No quantified TAM, but utility business will be a 'massive growth driver' and will roll out across multiple jurisdictions.** Utilities can seek to include Bloom servers in rate base as T&D assets deployed in-front-of-the-meter and procure per IRP plans, without waiting for specific customer POs. The logic: centralized generation's T&D costs, bottlenecks, and timelines misfit demand, pushing customers behind-the-meter; utilities then lose customers and rate base and must raise prices on others. Regulators are gradually allowing utilities to proceed. **Q: How is CCS commercialization progressing?** A: **The AEP–Google 700 MW project is a flagship; CCS is viewed as a trillion-dollar-plus opportunity.** Combustion-based technologies lack economical, scalable CCS pathways today; Bloom is the only commercial solution that directly yields high-concentration CO₂. Gas-fired generation with CCS can deliver reliable power with no local air pollution, no water use, and near net-zero emissions, offering the fastest decarbonization path once infrastructure is ready. **Q: Status of 800V DC adoption?** A: **The question is not 'if' but 'how fast'; timing depends on 800V DC server rack rollouts.** For a 1 GW data center, the company estimates over $5 bn in savings during the first five years of operations. The company is in talks with all hyperscalers, with strong interest; exact timelines are customer-driven, and customers have more say. **Q: How about CHP opportunities?** A: **CHP is a major growth vector for C&I customers and can be layered onto all power solutions.** On-site waste heat can be used for cooling, reducing power demand by ~20%; most industrial applications require steam or cooling. Whether AC or 800V DC, with or without CCS, CHP can be added on the same platform as needed. **Q: With demand scaling, can delivery cycles still stay within 90 days? How many MW in Nebius's first batch?** A: **No direct response on cycle time, and no MW disclosed for the first batch; emphasis is that the company will not be a bottleneck for a 300 MW customer.** Servers are ready for Nebius; data centers are built block by block, the first block is imminent, and the company is in place. Each subsequent block will be pre-staged, and other timing will be disclosed by the customer. Management cited Jensen Huang at GTC Berlin: a 1 GW data center loses about $10 bn of revenue per quarter of delay, i.e., 100 MW corresponds to $1 bn. That value outweighs other power-side considerations; Bloom is the cleanest for air permits and the fastest to energize among on-site options. **Q: Can automation be increased during expansion? How much cost reduction?** A: **No quantified cost-down; automation efficiency illustrated by flat headcount per line.** When stack output was 100 MW/year, about 600 employees ran the line around the clock; at 1 GW, the line still had ~600, a 10x efficiency gain. At 5 GW, headcount remains ~600; the company also benefits from AI across automation and other areas. **Q: For data centers, how big is the gap between LCOE and actual delivered power cost?** A: **Look at per-token power cost, not LCOE; on a full-cost basis, Bloom is lowest.** The cost chain runs LCOE → delivered grid price → rack power cost → token power cost: grid reliability is only 99.9%, requiring UPS for power quality and voltage/DC conversions. Grids and gas turbines cannot ramp quickly; racks must consume idle power to protect turbines, engines, or batteries, while fuel cells do not. No comment on the specific claim of 9% savings for a 1 GW facility. **Q: Water-use advantage vs. alternatives?** A: **About 1.4 mn gallons over the full lifecycle for 1 GW, vs. ~25 bn gallons for wet-cooled CCGT.** Management said the company's water use is smaller than rounding error to the second decimal place. **Q: How do you view the expansion cadence (market chatter: 250–300 MW per quarter)? Any supply constraints near 5 GW?** A: **No quarterly or annual expansion figures; capacity is tuned to demand, with no near-term signs of a step-down.** The extent of expansion fully depends on load demand; no response on supply constraints. **Q: With NIMBY pushing some developers overseas, how will the company capture that?** A: **Intl markets are attractive; overseas expansion is advancing, with updates likely soon.** In markets where gas prices are far above the U.S., Bloom's lower fuel consumption per token becomes an even bigger advantage. **Q: Largest single-site deployment in the U.S.? How does it compare with Wyoming and New Jersey?** A: **There are 100 MW-class projects; the Wyoming project is 700 MW, close to GW-scale.** The smallest system is 150 kW, used to power small Home Depot stores. **Q: Backlog is growing faster than revenue; how should we think about its growth framework?** A: **No directional range offered; clear guidance will be provided on the Q4 call.** Management only said backlog growth is 'steady'. **Q: Post AEP–Google, how interested are other hyperscalers in CCS? Impact on delivered power price?** A: **All hyperscalers are inquiring about CCS; the company can deliver an end-to-end solution within a 10%–15% cost uplift.** Management expects nearly all customers to procure at least 'CCS-ready' systems soon and connect once local CO₂ pipelines and storage wells are ready. They judge a 10%–15% uplift as a level most hyperscalers are willing to absorb voluntarily for zero-carbon power without mandates, and the company can cover full capture and storage costs within that band. **Q: How will contract assets convert to invoicing and collections?** A: **Contract assets may continue to grow; the key is to keep cycling to cash per billing milestones.** The CFO expects working capital to remain well-managed as the business scales, with AR and contract asset turns the top focus. **Q: How will CCS be monetized—by selling servers or via new service offerings?** A: **Monetization model-agnostic, offering multiple options based on end-customer needs.** Three paths: customers with storage capability (e.g., O&G) add CCS and monetize via carbon credits and power sales; partner for end-to-end solutions and revenue share; or customers retain carbon credits while paying the company via implied power price or equipment price plus margin. \ **Risk disclosure and statement:**[**Dolphin Research Disclaimer and General Disclosure**](https://support.longbridge.global/topics/misc/dolphin-disclaimer) ### Related Stocks - [AEP.US](https://longbridge.com/en/quote/AEP.US.md) - [ORCL.US](https://longbridge.com/en/quote/ORCL.US.md) - [GOOGL.US](https://longbridge.com/en/quote/GOOGL.US.md) - [GOOG.US](https://longbridge.com/en/quote/GOOG.US.md) - [BE.US](https://longbridge.com/en/quote/BE.US.md) - [ORCL-D.US](https://longbridge.com/en/quote/ORCL-D.US.md) - [GOOGN.US](https://longbridge.com/en/quote/GOOGN.US.md) - [ORCX.US](https://longbridge.com/en/quote/ORCX.US.md) - [ORCU.US](https://longbridge.com/en/quote/ORCU.US.md) - [ORCS.US](https://longbridge.com/en/quote/ORCS.US.md) --- > **Disclaimer: This article is for reference only and does not constitute any investment advice.**