Bloom Energy's Backlog Is 5 Times This Year's Expected Revenue. Here's How Fast It Can Actually Build.

Bloom Energy (NYSE:BE) disclosed a total backlog of about $20 billion in February -- about five times the revenue its raised 2026 guidance calls for. The fuel cell maker's systems generate electricity on-site, sparing a data center the yearslong wait for a grid connection, and that pitch has filled the order book.

But an order book rewards shareholders only as it converts into revenue. And the pace of conversion depends on what the backlog contains, on how fast Bloom can build -- and on when customers' sites are ready. Here's a closer look.

Image source: The Motley Fool.

A $20 billion order book

The figure arrived with Bloom's fourth-quarter report in February. About $6 billion of the total was product backlog (orders for the power systems themselves), a figure that had more than doubled in a year.

The rest, roughly $14 billion, was service: long-term operation and maintenance contracts attached to systems Bloom has installed or agreed to deliver. Those contracts run 5 to 20 years, and customers can cancel on an annual basis.

The company's definitions are worth noting, too. Product backlog counts existing contractual commitments for future purchases, and the dollar figure includes the tax incentives Bloom expects those deals to generate -- not just the revenue the company itself would book.

And the order book has kept growing since. In April, Oracle expanded its partnership with Bloom to cover up to 2.8 gigawatts of fuel cell systems for its artificial intelligence (AI) infrastructure build-out, with 1.2 gigawatts contracted up front and deployments continuing into 2027.

How fast can Bloom build?

The company's answer is a bigger factory. Bloom plans to double annual production capacity at its Fremont, California, plant from 1 gigawatt to 2 gigawatts by the end of 2026. Even then, the Oracle agreement alone could take more than a year of the expanded plant's entire output.

The conversion is already visible in the income statement. Revenue came in around $751 million in the first quarter, growth of 130% year over year. The second quarter brought $1.07 billion, up 166% year over year and 42% from the first quarter, with product revenue more than tripling to about $935 million. And management raised its full-year outlook twice -- in April to $3.4 billion to $3.8 billion, and in July to $3.9 billion to $4.2 billion, roughly double last year's total at the midpoint. Growth, in other words, is accelerating.

Set against about $4 billion of revenue a year (the guidance midpoint), the $6 billion product book is less than two years of work, and the factory could clear it faster as capacity doubles.

The service book pays out over decades

The service book is the reason I wouldn't read the $20 billion as a five-year revenue queue. That roughly $14 billion converts over the life of contracts running 5 to 20 years, and Bloom's service revenue in the first half of this year was about $131 million.

Of course, service revenue will likely keep growing as the installed base grows. But a book that size is built to pay out over decades, not a couple of build cycles.

There's also a gap, however, between the backlog and the much smaller figure on Bloom's books. Under standard revenue accounting, the company reported unsatisfied performance obligations of about $442 million for products and installation as of June 30, plus about $52 million for service. Bloom expects the product portion to become revenue within one to two years, in line with customers' project deployment schedules.

In other words, the backlog is the broad measure of what customers have committed to over time. The accounting figure captures only part of the work that is next in line.

The conversion timeline, then, splits in two. The product book is one to two years of work at the current build rate. The service book pays out over decades behind it.

The growth stock trades near $277 as of this writing.

That puts the stock's forward price-to-earnings ratio (today's price against next year's expected earnings) near 56, and the price at about 100 times the middle of management's own non-GAAP (adjusted) earnings per share guidance for this year.

A valuation that high arguably assumes full-speed conversion for years to come. Yes, the factory may well deliver. After all, management has already raised its outlook twice this year. But I'd rather not pay for those years up front -- at least not pay a price this high.

Daniel Sparks and his clients do not have positions in any of the stocks mentioned. The Motley Fool has positions in and recommends Bloom Energy and Oracle. The Motley Fool has a disclosure policy.