Bloom Energy (BE) just did the stock-market version of showing up to a knife fight with a battery pack and a smug grin. In a market still side-eyeing AI valuations, the fuel-cell power supplier posted numbers that made Nvidia’s growth look merely rich instead of mythical. The result: one of the louder proof points yet that the AI boom is not just about chips, but about the electricity needed to keep the whole circus spinning.
Sector heads can argue about hype, circular financing, and whether hyperscalers are building castles in the clouds. Meanwhile, the market is rewarding the companies that keep the lights on. Bloom’s quarter, plus the guidance lift that followed, turned it into a live case study in where the AI trade may be finding its next paycheck.
Bloom reported Q2 2026 results on July 28 and came out swinging. Revenue jumped 166% year over year to a record $1.07 billion, while product sales climbed 215.4% to $935.4 million as AI data-center operators rushed for onsite electricity. Adjusted EPS hit $0.78, nearly double the $0.41 analyst consensus, and non-GAAP gross margin expanded 604 basis points to 34.3%. The trading read: a power-infrastructure name with AI exposure, not a flashy AI software darling. The takeaway: when the market is desperate for compute and even more desperate for power, the picks-and-shovels crowd starts looking less like a side quest and more like the whole game.
Nvidia was the comparison point Bloom used to make its own growth look feral. Nvidia’s latest quarter showed 85% revenue growth to $81.6 billion, and the company remains the benchmark for the AI trade even as investors worry about hyperscaler spending and risk-reward math. But the stock itself has been acting more like a tired heavyweight than an invincible emperor, with a 7.4% five-day decline and just a 0.7% monthly gain in the latest cited moves. Trading profile: still the center of the AI market, but no longer immune to second-guessing. Investor takeaway: Nvidia is not broken, but the market is increasingly asking who else gets paid when AI expands beyond chips.
Oracle enters this story as the kind of customer that turns sales decks into receipts. Bloom’s CEO KR Sridhar said, “Bloom will not be your bottleneck,” after supplying Oracle’s data center within 55 days, and also said, “We deliver power at AI speed.” That matters because the whole AI infrastructure trade depends on proving that demand is not theoretical. Trading profile: not a move-sheet name in this pack, but a heavyweight whose data-center footprint gives Bloom’s pitch more credibility than a thousand buzzword-laced investor calls. The takeaway: when a major platform name needs power fast, the suppliers of fast power stop looking niche.
Tesla shows up here less as a direct Bloom competitor and more as proof that AI’s appetite for energy has escaped the server room. CEO Elon Musk said in a January 2026 conversation with BlackRock’s Larry Fink at the World Economic Forum in Davos, “The lowest-cost place to put AI will be space. And that will be true within 2 years, maybe 3 at latest.” That is not exactly a calm endorsement of the current grid. Trading profile: still an EV and autonomy stock, but also a megaphone for the idea that AI power demand is forcing investors to think bigger, stranger, and more electrical. Takeaway: once people start discussing space as a cheaper data-center location, the energy supply chain is no longer a footnote.
5. Bloom’s peer set in energy infrastructure: the market is chasing reliability, not romance
Bloom says its modular systems can be deployed in as little as 90 days, scaled through 325-kilowatt building blocks, and deliver up to 99.999% availability. That is the kind of language Wall Street loves when patience runs thin and grid connections take forever. The U.S. Department of Energy estimates data centers consumed 4.4% of U.S. electricity in 2023, with that share expected to reach 6.7% to 12% by 2028. Trading profile: not a single ticker, but the broader category of energy infrastructure tied to data-center buildout is getting the market’s attention because it solves a real constraint instead of selling a dream. Investor takeaway: AI needs chips to think, but it needs power to exist, and that makes the utility-adjacent names more interesting than they used to be.
Bloom’s profit engine was the part that made the quarter more than a one-night wonder. Non-GAAP operating margin jumped from 7.1% to 22.5%, and operating cash flow swung from a $213.1 million outflow to a $226.4 million inflow. CFO Simon Edwards added another line for the bulls to chew on: “Revenue grew 166%, while operating expenses grew just 48%.” Translation: the business is still growing fast, but it is not doing it with the financial discipline of a drunk guy using a corporate card in Vegas. For investors, that combination is what separates a true rerating candidate from a one-quarter fireworks display.
The guidance bump is what kept the market from filing Bloom under “cute story, no follow-through.” The company raised 2026 revenue guidance from $3.4 billion to $3.8 billion to $3.9 billion to $4.2 billion, implying 100% growth at the midpoint. Its adjusted EPS midpoint rose roughly 32% to $2.70, while operating-income guidance increased to $800 million to $900 million. In a sector where expectation management often feels like hostage negotiation, that is a big enough upgrade to make even jaded traders sit up and check the tape twice.
Bloom has also been living two lives at once. Year to date, the stock is up 100.5%, compared with Nvidia’s roughly 5.24% gain. But it has also dropped 23% in one week and 31% in one month, which is the market’s way of reminding everyone that a good story can still get punched in the mouth. That kind of whiplash usually means people are either taking profits, reworking valuations, or both. With Bloom, the selloff looks more like a pause after a violent run than a rejection of the business itself.
Then there is the valuation, because of course there is. Bloom trades at over 76 times forward non-GAAP earnings, which is 255% above the sector median, while Nvidia trades at 22 times forward non-GAAP earnings. That does not make Bloom cheap in any universe where the laws of arithmetic still apply. Across 26 analysts, Bloom’s average price target is about $286.20, with estimates ranging from $70 to $390. That spread tells you the market is still arguing with itself about whether this is an infrastructure winner or a momentum name wearing steel-toed boots.
The cleanest read here is not that Bloom is the next Nvidia. It is that AI is broadening into the power layer, and the companies that can supply electricity quickly, reliably, and at scale are getting paid for solving the problem everyone else is only now admitting exists. Bloom’s quarter showed real growth, real margin improvement, and real cash generation, which is more than most hype cycles can claim before they collapse into a crater.
For traders, the message is simple: Nvidia still runs the neighborhood, but Bloom just proved the utility bill can move markets too. If AI demand keeps outrunning grid capacity, the next leg of this trade may belong less to the chip vendor selling shovels and more to the company delivering power before the site contractor even finds a parking spot.
