Bloom Energy is no longer asking investors to value only a distant data-center opportunity. The company reported record second-quarter 2026 revenue of $1.065 billion, up 166% year over year, and raised full-year revenue guidance to $3.9 billion to $4.2 billion. At the midpoint, that would be roughly double 2025 revenue.
The growth case rests on Bloom’s solid-oxide fuel-cell systems delivering electricity on site, avoiding some of the delays associated with new transmission. That is attractive to data centers that need dependable power before a grid upgrade can be completed. It also broadens Bloom’s addressable market beyond emergency backup: the system can serve as primary distributed generation.
Growth quality matters more after a step-change
Bloom said second-quarter product revenue rose 215% and that the quarter delivered positive operating cash flow. Those are stronger signals than an order announcement because they show equipment shipment and cash generation in the same period. The comparison base was also meaningful: second-quarter 2025 revenue was $401.2 million, so the increase represents more than $660 million of additional quarterly sales.
Still, a record shipment quarter can create volatility. Fuel-cell projects require equipment manufacturing, site work and customer acceptance. Revenue may be concentrated in a few large deployments, and working capital can swing as inventory is built ahead of delivery. A large annual guide does not guarantee smooth quarterly progression.
The economics investors need to separate
Bloom’s product margin reflects manufacturing scale and pricing. Its longer-run economics also include installation, service obligations and the cost of financing customer projects. The company has identified upfront costs, interconnection and installation delays, service-related obligations and the possibility that backlog does not convert to revenue as material risks.
This distinction is especially important in data centers. Financing discussions or a proposed campus do not create booked supplier revenue. A signed equipment contract still may recognize revenue over milestones, while service commitments can continue for years. The best evidence is a combination of contracted backlog, delivery timing, gross margin and operating cash flow.
Scenario sensitivity
At the midpoint of the new guide, $4.05 billion, each one percentage point of gross margin equals about $40.5 million of annual gross profit. That BTI calculation shows why manufacturing execution can matter almost as much as the headline revenue number. A small change in yields, pricing or mix can move profit substantially when the revenue base doubles.
The broader power-buildout thesis is credible: constrained grids and rising data-center loads increase the value of generation that can be deployed on site. Bloom has now supplied evidence of scale rather than only potential. The next test is whether it can repeat positive operating cash flow while delivering the raised guide without sacrificing margin or accumulating costly service obligations.
The guide also changes expectations. With revenue already projected to double at the midpoint, merely meeting a high growth rate may not be enough if investors have capitalized several years of expansion. The valuation sensitivity is asymmetric: a project delay can remove revenue from a quarter while the factory and operating costs remain.
Capacity constraints elsewhere in the power system are a genuine tailwind, but they can shift rather than disappear. A fuel-cell installation still needs natural-gas supply, permits, equipment and site work. It can reduce dependence on transmission construction without eliminating infrastructure execution. In markets with carbon constraints, customers also have to weigh emissions attributes and any planned use of lower-carbon fuels.
The strongest evidence would pair a rising backlog with customer diversification. Concentration in a few large buyers can create rapid growth and abrupt volatility. A broader base across data centers and commercial users would make the raised guide more repeatable and reduce the consequence of any single campus delay.