Bloom Energy and Brookfield Expand AI Power Partnership Fivefold to $25 Billion

When Bloom Energy and Brookfield first agreed to build and finance fuel-cell power for artificial-intelligence data centers in October 2025, they called the $5 billion framework a first phase. Eight months later, they have shown just how much bigger the ambition was. On June 30, the two companies said Brookfield is expanding the financing framework fivefold, to $25 billion, to accelerate deployment of Bloom's on-site power systems for hyperscalers and AI developers racing to bring compute online faster than the electric grid can accommodate them.

The expansion, announced after the market closed on Tuesday, sent Bloom Energy (NYSE: BE) shares sharply higher, with the stock climbing roughly 12% in extended trading. It also crystallized a shift that has been building across the AI buildout for the better part of a year: the contest to power new data centers has become as much a financing race as an engineering one, and increasingly, capital and electrons are being bundled together from day one.

From $5 billion to $25 billion

The mechanics of the deal are straightforward, even if the scale is not. Brookfield supplies the capital and AI-infrastructure development muscle; Bloom supplies its solid-oxide fuel cells, which generate electricity on-site from natural gas or hydrogen without combustion. The framework funds those deployments globally, and the companies say the point is not simply to bankroll individual installations but to compress timelines by financing power, compute, and data-center infrastructure as an integrated package.

"When we formed this partnership, we said it was the first phase of a much larger vision," said Aman Joshi, Chief Commercial Officer of Bloom Energy. "Today's commitment reflects the momentum we are seeing in the market, as evidenced by recently announced large-scale deals. Bloom is uniquely positioned to address the urgent need for clean, reliable power to support the rapid growth of AI."

For Brookfield, the move slots into a far larger strategy. The expanded partnership sits within the firm's dedicated AI Infrastructure Fund, launched in November 2025 with a target to deploy $100 billion across AI factories, power solutions, and compute. Brookfield says it has already invested more than $100 billion in digital infrastructure and clean-power assets.

"Scaling our commitment with Bloom Energy reflects both the strength of this partnership and the conviction behind our broader AI infrastructure strategy, including integrated compute," said Sikander Rashid, Head of AI Infrastructure at Brookfield. "Scaling this partnership further strengthens Brookfield's position as one of the leading global AI infrastructure investors, capable of delivering end-to-end solutions, from electrons to tokens, for some of the world's most sophisticated customers."

Why on-site power is winning against the grid

The reason a fuel-cell financing framework can quintuple in eight months comes down to a bottleneck. As AI and cloud workloads explode, the constraint on new data centers is no longer chips or capital, it is dependable megawatts delivered on a schedule. Utility interconnection queues in major markets now stretch for years, and large-load requests are piling up faster than grid operators can process them. A campus that has secured GPUs, land, and financing can still sit idle waiting for a substation.

On-site generation sidesteps that queue. Bloom's fuel cells can be installed in months rather than the multi-year horizon of a new transmission connection, letting developers energize a site while a utility interconnection is still pending. That speed advantage is what has moved behind-the-meter power from a backup consideration to a primary tool for delivering AI capacity on time.

Analysts see the Brookfield-Bloom structure as a signal of something broader than a bigger check. "I don't think behind-the-meter power should automatically be viewed as a new asset class in isolation," Neil Osnato, founder of Persistence Analytics Group, told Data Center Knowledge. "Rather, it is part of a broader transition where energy certainty becomes a financeable asset. Investors are increasingly allocating capital not just to servers and buildings, but to the ability to deliver dependable megawatts on schedule when the grid cannot."

That framing captures why this deal matters beyond the two companies. By pairing financing with guaranteed power delivery from the outset, Brookfield is effectively selling schedule assurance, not just equipment. For operators, access to the capital that funds on-site generation may prove as decisive as access to the generation technology itself.

An energy race with many contestants

Bloom and Brookfield are not alone in chasing dedicated AI power, and fuel cells are only one of several competing bets. GE Vernova has pushed gas turbines for AI campuses, most visibly in the Chevron-Microsoft power partnership in West Texas. Wartsila is behind a 790-megawatt off-grid Texas data-center project. Battery storage, small modular nuclear, and renewables paired with firming capacity are all jockeying for the same load. Bloom itself has momentum beyond Brookfield: Oracle has agreed to acquire up to 2.8 gigawatts of the company's fuel cells, including 1.2 gigawatts already contracted.

What distinguishes the Brookfield-Bloom approach is the decision to embed financing into the generation model, making capital a lever for on-time delivery rather than merely a means of buying hardware. Whether that model becomes the template or simply an early example remains an open question.

What to watch

The next phase will be less about proving on-site generation works than about validating the assumptions underneath it. Watch whether the projected AI loads that justify these investments actually materialize, whether fuel supplies and operating costs stay sustainable at scale, and how behind-the-meter systems interact with utility planning as regulators such as FERC weigh new rules for large loads. The unresolved question is who ultimately bears stranded-asset risk if AI demand shifts or a site's compute plans change.

For now, the direction of travel is clear. As AI developers compete for dependable megawatts as aggressively as they once competed for GPUs, energy certainty is emerging as infrastructure in its own right, and the firms that can supply both the power and the capital to build it are positioning to win.