Bloom Energy's $25 Billion Brookfield Deal Splits the Fuel Cell Trade
Bloom Energy lined up $25 billion of Brookfield financing for data center power and carries a $25 billion backlog. Plug Power is walking away from the same demand.
Bloom Energy has secured a $25 billion financing arrangement with Brookfield to build AI-related power infrastructure for data centers and now carries a $25 billion backlog, while Plug Power is passing on the data center power opportunity.
Two companies that spent a decade being lumped together as "the fuel cell stocks" have now made opposite decisions about the single largest source of new electricity demand in the United States. Bloom Energy (BE) has locked in a $25 billion financing arrangement with Brookfield to build AI-related power infrastructure for data centers, and reports a backlog of the same magnitude. Plug Power (PLUG) is passing on that opportunity.
That divergence is the story, and it is more consequential than any quarterly print either company will publish this year. One firm is converting an electricity shortage into contracted, financed, multi-year work. The other has decided the economics do not justify chasing it.
What a $25 billion financing partner actually buys
The important word in the Bloom announcement is not "billion." It is "financing." Power equipment for hyperscale data centers is enormously capital-hungry, and the constraint on a company Bloom's size has never been demand — it has been the balance sheet required to sit between a manufacturing line and a customer that wants megawatts delivered on a schedule.
Bringing in Brookfield, an infrastructure investor whose business is precisely the long-duration financing of physical assets, changes the shape of that problem. It allows projects to be capitalized off a partner's balance sheet rather than funded by repeated equity issuance from Bloom itself — historically the mechanism by which shareholders in capital-intensive clean-energy names got diluted while revenue grew.
A $25 billion backlog alongside a $25 billion financing facility suggests a company trying to match the size of its order book to the size of its funding, rather than announcing wins it cannot bankroll. Whether the backlog converts to revenue on the timeline investors assume is the open question, and backlogs in energy infrastructure are notoriously sensitive to permitting, interconnection queues and customer schedule slippage.
Why Plug Power is saying no to the same demand
Plug Power's decision to pass looks strange only until you consider what its hydrogen business actually requires. Fuel cells that run on hydrogen need hydrogen — produced, liquefied, trucked and stored — and the cost of that supply chain is the central unresolved problem in the company's model. Data center operators buying primary or bridge power are ruthless about cost per megawatt-hour and about reliability of fuel supply. A hydrogen-fed solution has to clear both bars against natural gas, grid interconnection and, increasingly, on-site generation of every other kind.
Declining an opportunity you cannot serve profitably is a defensible act of discipline. It is also, from a market-narrative standpoint, expensive. Investors have spent two years rewarding any credible claim on AI power demand. A company that publicly steps aside from that demand is choosing balance-sheet realism over the multiple the story would have carried. The detail was reported by Nasdaq Markets.
The practical consequence is that Plug Power's path back to viable economics now runs entirely through its existing markets — material handling, industrial hydrogen, and whatever cost reductions it can extract from electrolyzer manufacturing. Those are slower-growing end markets than AI data centers, and they do not come with an infrastructure investor offering to write the checks.
How the share prices frame the two bets
The market is pricing these businesses as entirely different animals. Bloom Energy last traded at 265.63, down 5.39% on the day, after a session that ranged from 265.00 to 285.10 — a slide of roughly 7% from the intraday high to the close, illustrative of how violently a stock with this kind of embedded expectation can move on no company-specific news. Plug Power finished at 2.09, down 1.42%, in a range of 2.02 to 2.17. All figures are as of the last trade at 20:00 GMT on 18 September 2026.
The backdrop was mixed rather than hostile: the Nasdaq 100 (QQQ) closed at $721.45, up 0.63%, while the S&P 500 (SPY) slipped 0.12% to $761.69 and the Dow 30 (DIA) fell 0.48% to $515.88. Bloom therefore underperformed the tech-heavy benchmark by roughly six percentage points on the session — the kind of gap that tends to reflect profit-taking in a crowded position rather than a change in the thesis.
The backdrop was mixed rather than hostile: the Nasdaq 100 (QQQ) closed at $721.
A single-digit share price in Plug Power's case tells its own story about how much of the prior hydrogen narrative has already been unwound. Stocks at that level are typically being valued on survival and financing risk rather than on the terminal size of the addressable market.
The capital cycle behind both decisions
Strip away the technology and this is a familiar capital-allocation contest. AI data center construction is generating demand for on-site power faster than utilities can deliver interconnections, and that gap is being filled by whoever can supply generation at an acceptable cost and, crucially, finance it. The winners in that setup are rarely the companies with the best technology in the abstract. They are the ones whose unit economics work at scale and who can attract a partner willing to fund construction.
Bloom has now demonstrated both — at least on paper. Plug has concluded it can demonstrate neither in this end market, and has redirected accordingly. Both statements can be true simultaneously, and investors who treat the two firms as interchangeable proxies for the same theme are likely to be surprised in one direction or the other.
What to watch from here
For Bloom, the metrics that matter are conversion and margin: how much of the $25 billion backlog turns into recognized revenue, over what period, and at what gross margin once financing costs are allocated. A backlog is a promise, not a cash flow. Watch also for how much of the Brookfield structure sits off Bloom's own balance sheet, because that determines whether growth arrives with or without dilution.
For Plug, the test is narrower and more urgent: whether the core business can reach cash breakeven without the growth optionality that a data center pipeline would have provided. Passing on the AI power boom buys credibility with creditors only if the remaining business actually improves.
And for the wider sector, the split is a useful signal. When two companies with adjacent technologies look at identical demand and reach opposite conclusions, the differentiator is almost always the cost of the fuel and the cost of the capital — not the science.
Key facts
- Bloom Energy (BE) last close: 265.63, -5.39% (as of 18 Sep 2026, 20:00 GMT)
- Plug Power (PLUG) last close: 2.09, -1.42% (as of 18 Sep 2026, 20:00 GMT)
- Brookfield financing deal: $25 billion for AI-related data center power infrastructure
- Bloom Energy backlog: $25 billion
Frequently asked questions
What did Bloom Energy agree with Brookfield?
Bloom Energy secured a $25 billion financing deal with Brookfield to build AI-related power infrastructure for data centers. Brookfield is an infrastructure investor whose role is to fund long-lived physical assets, which allows large projects to be capitalized without Bloom carrying the entire construction cost on its own balance sheet.
How large is Bloom Energy's backlog?
Bloom Energy reports a $25 billion backlog, matching the scale of the Brookfield financing arrangement. A backlog represents contracted or committed future work rather than recognized revenue, so the key question for investors is how quickly and at what margin that figure converts into actual sales over the coming years.
Why is Plug Power passing on data center power work?
Plug Power is declining the data center power opportunity. Its fuel cell business depends on a hydrogen supply chain whose production, transport and storage costs must compete against natural gas and grid power on a cost-per-megawatt-hour basis. Choosing not to chase demand it cannot serve profitably is a balance-sheet discipline decision rather than a technology one.
Where did Bloom Energy and Plug Power shares last close?
As of the last trade at 20:00 GMT on 18 September 2026, Bloom Energy (BE) closed at 265.63, down 5.39% from a prior close of 280.76, within a day range of 265.00 to 285.10. Plug Power (PLUG) closed at 2.09, down 1.42% from 2.12, ranging between 2.02 and 2.17.
How did the broader market perform that session?
The Nasdaq 100 (QQQ) closed at $721.45, up 0.63%. The S&P 500 (SPY) eased 0.12% to $761.69, and the Dow 30 (DIA) fell 0.48% to $515.88. Bloom Energy's decline therefore ran well behind the tech-heavy benchmark, which suggests position-specific selling rather than a broad sector move.
What should investors watch next in these two stocks?
For Bloom Energy, the measures that matter are backlog conversion into recognized revenue, gross margin after financing costs, and how much of the Brookfield structure sits off Bloom's own balance sheet. For Plug Power, the test is whether its remaining markets can reach cash breakeven now that it has ruled out the data center growth avenue.
