Bloom Energy's Record Quarter Meets a De-Rating Tape

Revenue up 138%, guidance raised a third time — and the stock sits 37% off its June high. The pitch has quietly shifted from speed to rate base.

BE · Earnings Call · 2026-10-22

An accelerating business inside a de-rating tape

Bloom Energy's September quarter is, on the numbers, the cleanest print in its twenty-five year history. Revenue of $1.24 billion rose 138% year over year and 16% sequentially; product revenue crossed $1 billion for the first time at $1.07 billion, up 179%. “Operating margin for the quarter expanded to 23.1% from 8.9% a year ago.” — Simon Edwards · 2026-10-22 Guidance went up for the third time in 2026 — full-year revenue to $4.4–4.6 billion (roughly 122% growth at the midpoint), operating income to $960 million–$1.03 billion, EPS to $3.10–$3.40. Management began the year guiding 60% growth and is now guiding better than double that. And the stock is down. Bloom closed its recent 90-day window with a -22.5% return and sits 37% below its June 22 peak of $345.85, despite a full-history return of +771%. That gap is the story: the equity has been de-rating a business that keeps accelerating. The reason is visible in the wider tape. The AI data center theme, up 52% across a 360-day window with 41 advancers against 13 decliners, flipped to a -5.6% move over both the 90-day and 30-day windows — with 6 names up and 48 down. The market's AI-power basket has gone stale while Bloom's P&L has not. Even after the drawdown, valuation remains the elephant: price to trailing revenue of 28.5x through the June quarter, up roughly 737% year over year. A third guidance raise is not, by itself, enough to close that.

The vocabulary flipped: speed became rate base

What management talks about has changed, and it is the most underappreciated signal in the call. For six quarters the pitch was velocity — 90-day delivery, 55 days to first power at Oracle. Time to power, the company's signature phrase, has now fallen out of the curated top-30 entirely and registers as one of the sharpest conceptual decliners in Bloom's keyword set this quarter. So has Oracle-adjacent language — the marquee logo that launched the story is no longer the story, which matters given the project-finance wobble around Project Jupiter that Oracle has publicly flagged. In its place: grid power, distribution network, rate base, "solution of choice," electricity demand. Sridhar is now selling utilities on buying Bloom as regulated infrastructure rather than on beating a schedule: “the more they start seeing us as a distribution asset, through which they can reduce their total cost and still keep serving their customer. The better it is for the rate payer, the better it is for them to go to their policymakers and ask them to rate base their procurement of Bloom Energy servers.” — K. R. Sridhar · 2026-10-22 Note the political framing — Virginia is name-checked as "home to most of the data centers in the world" and ground zero for community friction, with the governor's energy plan cited as validating non-combustion fuel cells. This is a maturation of an old argument, not a new one. A year ago the utility pitch was a speed pitch: “You can now buy our fuel cells very quickly and install it in front of the meter and offer it to your customer.” — K. Sridhar, Chief Executive Officer · 2025-10-28 The difference now is that it is being pitched as a procurement channel — IRPs, front-of-the-meter, distribution deferral — which enlarges the funnel but also lengthens the sales cycle and puts returns in the hands of regulators. And the CEO's rhetorical confidence about competitive displacement has softened into something more conciliatory from the earlier "bridge to a bridge" era: “Earlier in this conversation, they used to bring up the concept of Bridge power with us. And I would smile and always say we're happy to sell you a bridge to a bridge because Superman coming.” — K. Sridhar, Executive, likely CEO or President · 2026-04-28 That swagger is gone; this quarter is about utility franchise politics, not turbine mockery.

Carbon capture stops being a someday slide — and 800V gets a number

The third quarter was another record quarter for Bloom. Revenue of $1.24 billion was up 138% year-over-year. And operating income of $286 million grew faster still, up over 500% year-over-year.
Simon Edwards · 2026-10-22

The option stack is where the substance sits. AEP disclosed a 700 MW carbon-capture-enabled deployment to deliver near-net-zero power to Google — a step change from earlier quarters where carbon capture sat near the bottom of Bloom's keyword list. It now ranks at the very top of the company's thematic set. Sridhar's framing: “when the world is ready for this, I see this as a greater than a trillion dollar opportunity.” — K. R. Sridhar · 2026-10-22 He put a price on the premium hyperscalers will tolerate — 10% to 15% for zero-carbon power — and said customers will soon be buying carbon capture-ready systems by default. The other lever is 800V DC, which also surfaced in the market-wide keyword set this quarter, meaning it is no longer a Bloom-only talking point. The claim is concrete: over $5 billion saved in a gigawatt data center's first five years by skipping the AC-DC conversion chain, plus roughly 20% of data center electricity avoided through combined heat and power. Whether or not one accepts the arithmetic, the strategic point lands — these are software-like add-ons to a hardware platform, which is why management insists Bloom is a platform and not a product.

The cash question, and a new CFO answering it

The quarter's one genuine blemish is working capital. Operating cash flow was a -$27 million use, free cash flow -$181 million, with capex of $153 million in the quarter and $231 million year to date. Cash still sits at $2.5 billion — roughly flat versus the start of the year despite investing four times what was spent in all of 2025. Year-to-date free cash flow remains positive at $42 million, and the June quarter itself printed $175 million of free cash flow, which is exactly why the swing looks worse than the trend. The mechanism is a keyword at the heart of the model: contract asset, now among Bloom's top-ranked themes and a direct target of analyst questioning. Product revenue is booked when control transfers; billing follows milestones. “While contract asset balance might grow, really important for us is that they are turning because reaching those billing milestones, and we are collecting cash.” — Simon Edwards · 2026-10-22 New CFO Simon Edwards — six months into the seat and on his second call — keeps returning to the same frame: “we already view this as a self funding business. So when you think about the operating performance of the business, that is what is funding the investments required to grow.” — Simon Edwards · 2026-10-22 That is the whole thesis in one line, and it is now falsifiable on a quarterly basis. The margin story is more mixed than the headline. Gross margin guidance was held at approximately 34% despite a higher revenue base — a mix and start-up cost issue, with product gross margin at 38.4% (up 250bp), services at 21.4% (up 700bp, second consecutive quarter above 20%), partly offset by installation revenue timing. The operating line, though, is the proof of the model: operating margin of 18.8% in the June quarter, now 23.1% in September, versus deeply negative levels for most of the last decade. Revenue has grown roughly tenfold on a nine-year view; capex through June was only about 5% of sales, consistent with management's capital-light claim.

What to watch

Bloom is asking investors to believe three things at once: that a 5 GW capacity base is a milestone rather than a destination (hence the Fremont footprint expansion, which analysts pressed on twice), that carbon capture and 800V are near-term revenue rather than 2030 optionality, and that working capital normalizes as billings catch up. The first is credible given backlog growing faster than revenue; the second now has a 700 MW anchor; the third is the one that determines whether a 28x revenue multiple survives the next two prints. The tape has already voted once — the AI power complex rolled over while this business compounded. The next quarter of backlog disclosure, and any international win management keeps teasing, is where the two re-couple or diverge for good.