Bloom Energy Lifts 2026 Revenue Guide to Over 120% Growth as Fuel Cells Move From Data Center Niche to Utility and Carbon Capture Platform
Third quarter earnings call, October 22, 2026: record $1.24 billion quarter, third guidance raise of the year, but operating cash flow turns negative
Bloom Energy reported third quarter revenue of $1.24 billion, up 138% year-over-year and 16% sequentially, and raised its full year outlook for the third time this year. Management now guides to revenue of $4.4 billion to $4.6 billion, roughly 122% growth at the midpoint. On the first quarter call, the company was expecting 60% growth. Year-to-date revenue of $3.05 billion is already 50% above all of 2025. The company also noted that it was added to the S&P 500 during the quarter.
Operating leverage is the real story in the numbers
Operating income of $286 million rose more than 500% year-over-year and exceeded the entire prior year's operating income in a single quarter. Operating margin expanded to 23.1% from 8.9% a year ago, as revenue grew 138% while operating expenses grew only 26% to $141 million. Non-GAAP EPS was $0.94, against $0.15 a year ago. Adjusted EBITDA was $299 million, or 24% of revenue.
Product revenue crossed $1 billion for the first time, reaching $1.07 billion, up 179% year-over-year and 87% of the total. Product gross margin was 38.4%, up 250 basis points year-over-year and 120 basis points sequentially, achieved while the company ramped new production lines and expedited deliveries. Services margin reached 21.4%, the second consecutive quarter above 20%. Total gross margin was 34.5%, held back by the timing of lower-margin installation revenue, which rose 66% sequentially to $84.9 million.
The updated guidance implies an operating income range of $960 million to $1.03 billion, a 22% margin at the midpoint versus 14% in the February outlook, and non-GAAP EPS of $3.10 to $3.40. One point deserves scrutiny: full year gross margin guidance was held at about 34% despite higher revenue. CFO Simon Edwards attributed this to near-term costs of new line start-ups and expediting, and noted that engineering and sourcing cost-reduction programs "will take some time to flow through to COGS." The message is that product-level pricing is strong, but consolidated margin upside is being traded for speed of delivery. Edwards declined to address 2027 directly.
Cash flow is the soft spot
Cash flow from operations was a use of $27 million, and free cash flow was a use of $181 million in the quarter, driven by growth in inventory, contract assets and receivables. Capital expenditures were $153 million in the quarter and $231 million year-to-date, roughly 4x all of 2025, for added capacity in the existing Fremont and Newark facilities. Year-to-date free cash flow remains positive at $42 million, and the company ended the quarter with $2.5 billion in cash, roughly flat versus the start of the year.
Management framed the shortfall as timing, with revenue recognized on transfer of control while billing follows contract milestones. Edwards said the quarter's cash use "is not the beginning of a trend" and described the business as self-funding. Still, with working capital absorbing cash at this pace, contract assets are a number to watch. Edwards said the focus is on ensuring that contract assets and receivables "are turning" as billing milestones are hit, but offered no figures on expected conversion.
Capacity expansion beyond 5 gigawatts, with no timeline
Asked about the recent decision to expand the Fremont real estate footprint, and whether capacity would move from the current 5 gigawatts of annual output to above 10 gigawatts, CEO K.R. Sridhar declined to give a number or timeline. He said the 5 gigawatt milestone "is just a milestone in a long journey" and that the company needs "to get prepared for what is after that." He described capacity as an "analog dial" rather than a step function: "We will never go past our headlights, but we will never be slow enough where we are the bottleneck to our customer getting the power." Management gave no quarterly or annual capacity increments, though he said he does not see the dial moving down anytime soon.
On automation, Sridhar noted that stack production once required about 600 employees for 100 megawatts a year, while a gigawatt of stack output will need about the same 600 employees, and he said the same headcount would remain even at 5 gigawatts.
Carbon capture emerges as a concrete, if early, opportunity
The most notable new commercial development was AEP's disclosure that 700 megawatts of carbon capture and sequestration-enabled Bloom Energy Servers will supply Google with near net zero electricity, which Sridhar called a first-of-its-kind project. He argued the technology is structurally advantaged: "Combustion technologies do not offer an affordable and scalable pathway to carbon capture today," because the fuel cell produces a concentrated CO2 stream. He put the long-term opportunity at "greater than $1 trillion," a figure that is aspirational and conditioned on "when the world is ready."
More practically, Sridhar said all hyperscale customers are asking about the option, and that "almost all our customers would very soon at a minimum start buying systems from us that are carbon capture ready." He cited 10% to 15% as a premium most hyperscalers would voluntarily accept for zero-carbon power, and said Bloom can deliver all-in carbon capture and sequestration within that range. On monetization, the company will be flexible, from bolt-on arrangements with gas companies that capture the credits, to partner-led end-to-end solutions, to customers paying for equipment plus a margin. That flexibility is sensible but leaves the economic model for investors undefined.
Utilities as a new front-of-the-meter channel
Sridhar signaled a strategic expansion toward utility-owned, front-of-the-meter deployments. His argument is that utilities facing customers moving behind the meter risk losing rate base and shifting costs onto remaining ratepayers. Placing fuel cells in the distribution network, rate-based and procured through integrated resource plans without a specific customer in mind, avoids transmission bottlenecks. "The utilities are finally getting the leeway they need from their regulators to include these assets, and so they're jumping on the opportunity," he said, adding that he expects significant wins. No utility contracts were quantified, so this remains a thesis rather than a reported backlog driver.
Time to power and the Nebius deployment
Bloom says it delivered power to Oracle's Utah site in under 2 months, and at Nebius in New Jersey obtained planning board approvals within 3 months of contract signing and had the first power block ready 2 months later. Sridhar said Bloom's systems are ready for Nebius's 300 megawatt data center and will be in place block by block, but declined to specify the size of the first block or customer timing. He leaned on a comment attributed to NVIDIA's CEO that each quarter of delay on a gigawatt data center costs $10 billion in missed revenue, arguing this "overwhelms every other value proposition" on the power block.
On scale, the largest deployments are in the 100 megawatt class, and the Wyoming project, at 700 megawatts, approaches gigawatt scale. The smallest systems are 150 kilowatts.
800V DC, combined heat and power, and the cost argument
Sridhar called 800V DC "a no-brainer," citing a company report that a 1 gigawatt data center would save over $5 billion in its first 5 years by eliminating power conversion steps. He said conversations are under way with all hyperscalers, but that adoption depends on the arrival of 800V DC server racks, and "they are in a better position to answer the timing question than I am." Combined heat and power, which he said can offset about 20% of power demand through cooling, is positioned as a lever for commercial and industrial customers. Both remain optionality rather than revenue.
Responding to UBS, Sridhar argued the right metric for data centers is cost per token rather than LCOE, tracing the chain from LCOE to delivered grid power to rack power to token power cost. On water, he said a wet-cooled combined cycle plant serving 1 gigawatt would consume roughly 25,000 million gallons over its life, versus 1.4 million for Bloom. These claims come from management and were not independently tested on the call.
Demand breadth and international
Management pointed to five commercial wins: a conditional U.S. Army award to partner Indelible for long-term leases on installations, the AEP carbon capture project, first shipments into Texas this quarter, utility relationships, and international interest, with partner discussions under way and announcements promised soon. Sridhar said the higher cost of natural gas abroad amplifies the efficiency advantage. A company survey of over 200 commercial and industrial businesses found that 25% of facilities expect to be fully powered on-site by 2030 and 35% by 2035, though this is self-commissioned data.
Backlog was described only as growing faster than revenue, and Sridhar declined to quantify it, promising full guidance with the next report. Given the pace of revenue growth, that qualitative claim sets a high bar, and the lack of a figure is a gap for investors trying to size 2027. Edwards also noted that the timing of "a small number of deliveries" around year-end will determine where in the revenue range the company lands.