Quantum Yields

**Disclosure:** I own shares of Bloom Energy (NYSE: BE). This is one of my core holdings, not a detached observer’s take. Nothing here is financial advice — do your own work. See the full disclaimer at the end.

Quick Thesis

Every conversation about AI eventually runs into the same wall: you can build the models, buy the GPUs, and sign the customers, but none of it turns on without electricity — and the grid can’t deliver it fast enough. Interconnection queues and gas turbine lead times now run years. Bloom Energy sells the workaround: solid-oxide fuel cells that generate power *on-site*, shipped in months instead of years. It’s a hardware sale with a decade-plus service contract stapled to the back of it, so recurring revenue compounds behind every box installed. The business is inflecting hard — Q1 2026 revenue grew 130% year-over-year, margins are climbing, and management raised full-year guidance meaningfully. The whole risk, in my view, isn’t the business. It’s the price.

Business Overview

Bloom makes what it calls “Energy Servers” — solid-oxide fuel cell systems that convert fuel (natural gas, biogas, or hydrogen) into electricity *electrochemically*, without combustion. In plain terms: it’s a box that sits on a customer’s site and makes clean-ish, concealable, reliable power without waiting for the utility to run new lines. Customers are data centers, semiconductor fabs, hospitals, utilities, and other commercial and industrial users who need power *now*.

The model has two parts, and the second is the one people underrate. First, Bloom sells the hardware. Then it attaches a long-term service contract — reported near-100% attach rate, terms running a decade or longer — that layers a recurring, higher-margin annuity on top of each install. So the installed base isn’t just a revenue event; it’s a growing book of contracted service income that compounds quietly in the background.

Founder KR Sridhar still runs the company, roughly 25 years in. The core technology traces back to work he did at the University of Arizona on oxygen generation for NASA’s Mars missions. I put real weight on founder-led companies, and a quarter-century of one person steering the ship through a brutally capital-intensive build-out is exactly the kind of continuity I look for.

The Numbers That Matter

Bloom’s Q1 2026 (reported late April) was the quarter that changed the conversation:

– Revenue of $751.1 million, up 130% year-over-year, driven by *product* revenue growth of 208%.

– Non-GAAP gross margin of 31.5%, up roughly 280 basis points year-over-year.

– Operating income of $129.7 million versus $13.2 million a year ago — operating margin climbing to 17.3%, a swing of more than 1,300 basis points.

– Adjusted EBITDA of $143 million, nearly a 19% margin.

– First positive Q1 operating cash flow in company history, and $2.52 billion of cash on the balance sheet.

And then management did the thing that matters most — they raised the full-year bar. FY2026 revenue guidance went from $3.1–3.3B up to $3.4–3.8B (roughly 80% growth at the midpoint, with the bottom of the new range sitting above the top of the old one). Full-year non-GAAP gross margin guidance was lifted to ~34%, up from ~30% in 2025.

Read those two lines again — revenue accelerating and margins expanding and guidance raised — and you have the textbook signature of operating leverage kicking in. That’s the entire bull case in one sentence: as volume scales, margins expand. Backlog now sits around $20B, roughly 10x the prior year’s revenue, which gives real forward visibility to that ramp.

The Drive: Bring Your Own Power

The thesis-defining product is the AI-data-center use case Bloom brands as “bring-your-own-power.” When grid interconnection is the bottleneck, a data-center developer can deploy Bloom instead of waiting on the utility. The proof point is Oracle’s “Project Jupiter” — a roughly 2.45 GW grid-independent AI campus in New Mexico that replaces planned gas turbines and diesel entirely with Bloom fuel cells. Add American Electric Power (~900 MW) and a $5 billion Brookfield financing partnership to deploy globally, and you can see the flywheel: reference wins beget more reference wins.

Capacity is scaling to match — 1 GW to 2 GW by end of 2026, with a manufacturing footprint that can support up to ~5 GW.

Moat

Moderate and improving, not yet ironclad — and I want to be honest about that. The moat comes from four places: (1) proprietary solid-oxide stack technology and ~25 years of accumulated engineering; (2) manufacturing scale no fuel-cell peer comes close to matching; (3) the service-contract annuity that locks in decade-plus customer relationships and creates switching costs; and (4) marquee reference wins (Oracle, AEP, Brookfield) that function as their own sales engine.

But notice what that moat is: scale, a head start, and embedded service — not “impossible to replicate.” That’s why I hold BE as a bet on company-building rather than as unassailable ballast. The real competition isn’t other fuel-cell makers (FuelCell Energy and the like are smaller and less proven); it’s the alternatives a data center could choose — gas turbines, diesel gensets, and eventually small modular nuclear. Bloom’s edge is speed-to-power, cleaner-than-combustion emissions, modularity, concealability, and reliability. Turbines currently have supply-chain queues of their own, are noisy, and generate greater emissions, which is helping Bloom win campus-scale deals outright. Jupiter is the proof.

Risks

Valuation is the whole risk. I’ll say it plainly because it’s the thing I actually lose sleep over. The business is executing beautifully — but the stock has run hard, up multiples in a year, and it trades at a rich forward multiple. When a stock is priced for perfection, any stumble gets punished severely. And there are several places a stumble could come from:

– Deployment timing. Revenue is lumpy. Big campus deals (Oracle, Nebius-type) can slip a quarter, and the market rarely forgives a miss on a name priced like this.

– The margin test. The entire premium rests on the “scale drives margin” thesis. The clean test is whether that ~34% gross-margin guide holds as revenue ramps toward the $3.4–3.8B target. Holding it confirms the whole story. A margin miss tied to deployment delays would be the first real crack. Q2 2026 (expected July 30) is the number to watch — non-GAAP gross margin against that ~34% guide.

– Fuel dependence. Economics still lean on natural gas as the primary fuel, which is a long-term decarbonization asterisk — though the “power now” argument overrides it for customers today.

– Concentration. A handful of very large campus deals drive a lot of the story.

Sustained Alpha

Bloom is selling the single scarcest input in the entire AI build-out — immediately available power — and it’s the only commercially scaled fuel-cell company with the manufacturing footprint to serve gigawatt demand. Q1 showed genuine operating leverage, the guidance raise showed management’s own confidence, and the founder is still at the wheel after 25 years. That’s a business I want to own for the long haul.

The discipline is in remembering that a great business and a great stock are two different questions. I’m not chasing this one on strength. The way I think about it: I’d rather add on weakness — a deployment-timing wobble or a broad-market pullback that takes a richly valued name down with it — than pay up when the crowd is euphoric. That’s the same principle I keep coming back to:

> “Be fearful when others are greedy and greedy when others are fearful.” — W. Buffett

Bloom has earned its place in my book. The price will decide the entry.

*Disclaimer: I own shares of BE. This post reflects my personal opinions and analysis for informational purposes only. It is not financial, investment, or tax advice, and it is not a recommendation to buy or sell any security. Figures are drawn from company filings and public reporting as of mid-July 2026 and may be out of date by the time you read this. Markets involve risk, including loss of principal. Do your own research and consult a licensed professional before investing.*