Bloom Energy’s Fast-Payback Factory Model Lets CEO Sidestep Equity Dilution

Bloom Energy chief KR Sridhar has thrown cold water on any expectation of a share sale, insisting the fuel-cell maker can bankroll its own breakneck expansion. With demand from AI data centers exploding, the company’s production lines are paying for themselves within six months — a math that makes a capital increase unnecessary in Sridhar’s view. For investors who had been bracing for dilution after the stock’s meteoric rise, the message was a relief.

The economics are straightforward: building a new gigawatt of fuel-cell manufacturing capacity costs Bloom between $100 million and $150 million. That investment is recouped in roughly half a year through sales of the units themselves. Sridhar points to that rapid payback as the engine behind a “self-financing” growth strategy, one that keeps the company off the capital markets even as it scales to meet the voracious power needs of hyperscale data centers.

That operational confidence is backed by a balance sheet that has flipped decisively into the black. In the first quarter of 2026, Bloom’s revenue surged 130% to $751 million, swinging from a loss of $23.8 million a year earlier to a net profit of $70.6 million. Service margins improved from 1.3% to 13.3%, and the company generated $73.6 million in operating cash flow — a performance that underpins the no-equity-raise stance.

Yet the stock’s staggering run means the market has already priced in plenty of good news. Bloom shares have climbed roughly 1,400% over the past twelve months and nearly tripled since the start of the year, closing recently at $273.51. The catalyst has been a string of billion-dollar infrastructure deals — most notably a partnership with Oracle worth up to 2.8 gigawatts and a $2.6 billion pact with Nebius Group. At the end of the fourth quarter of 2025, the product backlog stood at $20 billion, two and a half times the level of a year earlier.

Should investors sell immediately? Or is it worth buying Bloom Energy?

That backlog is not without potential speed bumps. BMO Capital Markets has flagged a risk around the Oracle Jupiter project in New Mexico, which calls for 2.45 GW of Bloom’s fuel cells. The concern centers on the Green-Chile Pipeline, a critical gas-supply link that may not have met all requirements for an accelerated permitting process at the Federal Energy Regulatory Commission. FERC staff reportedly flagged an April filing for missing documentation on protection of historic sites. BMO describes the situation as a headline risk for now and does not expect it to affect Bloom’s 2026 financial forecast; the bank maintains a “Market Perform” rating with a $279 price target.

Sridhar, meanwhile, is leaning into modularity and lower emissions versus gas turbines as selling points, particularly for AI data centers that need power quickly. The market for data centers is projected to grow from roughly $300 billion this year to nearly $700 billion by 2034, and Bloom aims to position its technology as the standard for critical AI infrastructure — all without tapping equity markets.

Insiders have taken advantage of the stock’s ascent to lighten their holdings. Over the past twelve months, corporate insiders have sold a net $83 million more in shares than they purchased. Board member John Chambers disclosed plans to sell 55,000 shares on May 28. The selling does not undermine the growth story, but it adds a note of caution for a stock trading at 147 times forward earnings estimates for fiscal 2026.

Analyst sentiment remains broadly positive, with 14 of 16 covering brokers voting “buy” and only two recommending a sale. The consensus price target stands at $263.13, while the highest estimate reaches $335. For the full year 2026, Bloom expects revenue between $3.4 billion and $3.8 billion — an 80% jump from the prior year, up from an earlier forecast of around 60% — and adjusted earnings per share of $1.85 to $2.25. The company’s ability to turn that growth into cash without diluting shareholders will be the next critical test.