Bloom Energy Stock After Q3 2026 Earnings: Revenue Up 138%, Operating Cash Flow Negative

Bloom Energy reported its third quarter after the close on October 22. Revenue was $1.24 billion, up 138%, product revenue passed $1 billion for the first time, and the company raised its 2026 revenue guidance for the third time this year, to $4.4 to $4.6 billion. In April the range was $3.4 to $3.8 billion. The next day the stock fell 11.2% to $193.30 on more than twice its normal volume.

The reason was one line in the cash flow statement. Operating cash flow was −$27 million in a quarter where Bloom earned $250 million. Add $153 million of capex and free cash flow was −$181 million. I think the selloff overstated how big the cash problem is, but it's pointing at something real about Bloom's position.

Before the detail, here's the chart I'll be updating every quarter. The grey line is what customers owe Bloom, revenue it has recognised plus invoices it hasn't collected. The gold line is cash customers have paid before delivery. A supplier that sets its own terms would have the gold line on top. Bloom's is below, and in Q3 the gap widened.

Most of the gap is contract assets. That's revenue Bloom has recognised because it delivered and installed equipment, but can't invoice yet because the contract's billing milestone hasn't been reached. They rose $430 million in the quarter, and current contract assets ended September at $793 million, up from $179 million at the start of the year. Receivables, which are invoiced but unpaid, rose another $175 million. Inventory rose $170 million, which CFO Simon Edwards said on the call was built "for future deliveries."

Edwards was direct about the mechanism: "cash collection can follow revenue recognition, with the difference reflected in contract assets and receivables" until billed and collected. Asked about it, he said "this quarter we used cash from operations, but that's not the beginning of a trend," and on contract assets, "what I do expect is they're going to turn." He also said Bloom expects "to continue investing in working capital to support future growth."

The capex half of this doesn't worry me much. It's machinery and production capacity at Fremont and Newark, Bloom has $2.49 billion of cash against $2.47 billion of recourse debt, and Edwards said the cash balance is roughly where it started the year. Nine months in, free cash flow is still positive at $42 million. A company that has doubled revenue and just raised guidance should be spending on capacity.

The working-capital half is the more interesting one, because it says something about who has the leverage. The pitch for every AI power stock this year has been scarcity. Data centres need power faster than the grid can deliver it, so whoever can supply it sets the terms, and I'd expect that to show up as customers paying before delivery.

That's what GE Vernova looks like. In Q3 2026 it reported $5.4 billion of free cash flow on $12.6 billion of revenue, and CFO Ken Parks said working capital was a $5.3 billion cash benefit, "driven primarily by down payments on increased orders." In Q2 the same line was $6.4 billion. Customers are paying to hold a place in a turbine queue that, with backlog plus slot reservations at 122 GW, runs years out. Bloom's customers take delivery first and pay when the milestones come due.

Here's my reading of the stock files, and it's my reading rather than anything the framework scores directly. Both companies are rated strong on transaction embedding, because once a turbine or a Bloom server is running the customer is locked into its service contract. The difference is before the sale. GE Vernova is rated strong on bundling and is credited with scale economics, and it sells turbines with multi-year lead times that buyers queue for. Bloom isn't credited with scale economics, its bundling is rated intact, and its permitting edge is regulatory arbitrage that narrows if turbine permitting is fast-tracked. Before a server is installed, Bloom is competing on speed to power against gas engines and turbines, and I'd expect that to show up as weaker payment terms, which is what the quarter showed.

At the October 23 close Bloom traded at roughly 59 times the midpoint of this year's guided non-GAAP earnings, and a multiple like that assumes the ramp keeps going on good terms. The cash flow statement says Bloom is a fast-growing manufacturer that funds its customers' timelines. I think those can coexist while revenue doubles, but if growth slows and the receivables don't, the second description is the one the stock will be priced on.

The strongest case against me is in the cash flow table. Deferred revenue and customer deposits, which is cash customers have paid before Bloom has delivered, rose $309 million in Q3 and $212 million in Q2. At September 30 they totalled $754 million, up from $88 million a year earlier. The non-current portion alone went from $43 million at the end of 2025 to $475 million, which looks like long-dated prepayments on large projects. If that keeps going, Bloom's customers are doing exactly what GE Vernova's do, and my cross-read is a snapshot of a company partway through changing its terms.

The second argument is that this has happened before and fixed itself. At the end of Q3 2025, contract assets and receivables were $671 million against $519 million of quarterly revenue, a higher ratio than this quarter's. In Q4 2025 those balances fell by $120 million and operating cash flow was $418 million. Bloom recognises revenue as it delivers on large multi-site projects and bills on milestones, so a big delivery quarter will always leave a hole that the next billing quarter fills. On that reading, Q3 2026 is the same pattern at twice the size, and the 11% drop was the market noticing a seasonal swing it had ignored when the numbers were smaller.

And a bull would say the GE Vernova comparison is unfair. Turbines take years to build, so a down payment is the norm in that industry whatever the demand. A Bloom server ships in months, and speed is the product. Asking a buyer to prepay for something that arrives in a quarter would give away the reason to choose Bloom. On that view the payment terms come with the product and say nothing about leverage.

I take all three seriously, and the honest answer is that this is a matter of degree. Prepayments are rising fast, and $475 million of non-current deposits says some customers are paying ahead on long projects. But they cover about 53% of what Bloom is owed, and the gap between the two widened to $673 million in Q3 from $379 million in Q2, while GE Vernova's working capital brought in $5.3 billion in a single quarter. Q4 2025 is a fair precedent for the owed balance falling, and I'd expect some of that again. On lead times, the bull has a point I can't fully answer with numbers: I don't have payment terms for a short-lead product like a gas genset to compare with. What I can say is that Bloom's own customers are prepaying more each quarter, which tells me the terms are moving, and they haven't moved far enough yet for Bloom to stop carrying $1.43 billion of their bills.

Bloom reports Q4 early next year. The first thing I'll check is whether the owed line in the chart at the top falls the way it did a year ago. If it does, Edwards was right that it was timing, and the quarter's cash scare goes away. If it doesn't, and it keeps growing faster than revenue, the ramp is being financed on Bloom's balance sheet as well as limited by its scandium supply, which the stock file already flags.

The second thing is the other line. If deposits and deferred revenue catch up with what Bloom is owed at any quarter-end through June 2027, customers are paying ahead, Bloom has the terms a scarce supplier gets, and I've misread the quarter. I'll extend this chart with each report.