Why Bloom Energy Took Over a Second Factory Before It Needed It

A month ago, Bloom Energy quietly took over a second building in Fremont, California: 158,000 square feet, nearly the same size as the factory it already runs there. On Thursday, the fuel-cell maker gave the market the quarter that explains the purchase: record revenue of $1.24 billion, up 138 percent from a year ago, and a 2026 forecast raised for the third time in nine months.

The market’s answer was a shrug. Revenue beat the $1.07 billion FactSet consensus by more than 15 percent, and adjusted earnings of 94 cents a share cleared the 69-cent estimate by more, but the stock fell 8.7 percent to $217.77 and kept sliding after the close. It now sits about 37 percent below its June record of $345.85. JL, a retail investor posting on Threads, wrote what the tape was thinking: “DUMPING ON A DOUBLE BEAT + RAISE.”

Read the guidance ladder and the shrug looks premature. Bloom opened the year guiding to $3.1–$3.3 billion in revenue, lifted that to $3.4–$3.8 billion in April, then to $3.9–$4.2 billion in July. Thursday’s number, $4.4–$4.6 billion, is the third raise. The arithmetic the release leaves to the reader: with $3.05 billion already booked through September, the new range implies a fourth quarter of $1.35–1.55 billion, as much as 25 percent above the quarter just reported. That would be a fourth straight record.

Chief financial officer Simon Edwards tied the raise to “year-to-date performance and customer demand,” adding that the company keeps investing “to build the capabilities and capacity needed to scale the business.” The capacity is concrete: Bloom is spending $100 million to double its Fremont plant’s annual output from 1 gigawatt to 2 by the end of 2026, and the second building in the same city suggests the doubling won’t be the last expansion. Almost all of it is hardware, with product revenue at 87 percent of the quarter’s total, and the order book disclosed in February stood near $20 billion, about $6 billion of it product.

Founder and chief executive KR Sridhar put where the demand is going in one sentence: “What began with hyperscalers is rapidly moving across market segments because every business needs reliable power to grow.” A month earlier, Ameren Missouri’s 20-year plan had called for 500 megawatts of natural-gas fuel cells by 2030 — supplier unnamed, plan nonbinding, but a utility planning fuel-cell fleets alongside the data-center orders. Brookfield Asset Management’s financing framework for Bloom deployments, meanwhile, has grown fivefold to $25 billion.

Not everything is converting on schedule. Oracle’s Project Jupiter, where Bloom is the sole power provider for up to 2.45 gigawatts in New Mexico, has seen a payment notice that could push out timing; Bloom has said the fuel cells could be sold elsewhere if needed, according to Insider Monkey’s reporting. The company’s own risk list names air-permitting delays and whether its backlog converts into revenue.

The elsewhere already exists. On October 14, American Electric Power said it would spend about $4 billion on a 702-megawatt fuel-cell project near Cheyenne, Wyoming, using Bloom’s technology under a 20-year power purchase agreement with Google. Construction is scheduled to start in 2026, with full operation expected by the end of 2028. The project company, Cheyenne Fuel Cell, LLC, is about 92 percent owned by AEP.