Brookfield Asset Management has significantly expanded its financing partnership with fuel cell manufacturer Bloom Energy, increasing its commitment from $5 billion to $25 billion. The agreement, originally established in October 2025, is aimed at funding onsite power infrastructure for AI data centers, with the expanded deal reportedly beginning with an initial site in Europe before scaling further globally.
The scale of the increase, a fivefold jump in under nine months, reflects how quickly capital is moving to address one of the AI industry’s most pressing constraints: getting enough reliable power to data centers fast enough to keep up with demand.
How the technology works
Bloom’s solid oxide fuel cells generate electricity onsite through an electrochemical reaction between hydrogen and oxygen, avoiding combustion entirely. Because they don’t rely on new transmission infrastructure or grid connections, they can be deployed considerably faster than traditional power buildouts, and the companies describe them as cleaner than diesel generators or gas turbines often used as stopgap power sources.
Bloom has already installed hundreds of megawatts of this technology at data center sites, working with partners that include American Electric Power, Equinix and Oracle. Bloom’s headquartered in California, employs roughly 2,300 people and produces about 1 gigawatt of fuel cell capacity annually across nine countries.
Part of a much larger AI infrastructure push
The expanded commitment sits within Brookfield’s dedicated AI Infrastructure Fund, launched in November 2025 with a target of deploying $100 billion. Brookfield already manages more than $1 trillion in assets and has invested over $100 billion in digital infrastructure and clean power to date.
Executives from both companies described the expansion as reflecting growing momentum in the market. Bloom’s leadership pointed to the urgent need for reliable power to support AI’s rapid growth, while Brookfield’s infrastructure team framed the deal as part of a broader strategy spanning both power generation and computing capacity.
A market moving faster than the grid can keep up
The rapid scale-up says as much about the state of the broader power grid as it does about Bloom’s specific technology. Utilities in major markets have struggled to keep pace with the electricity demands of large scale computing, pushing developers toward behind-the-meter solutions that bypass lengthy queues for new grid connections. That dynamic has turned onsite power generation from a niche engineering choice into a strategic necessity for some of the world’s largest technology companies.
Fuel cells still face competition from gas turbines, which continue to dominate the behind-the-meter power market on cost grounds, and carry higher upfront capital costs even though their operating expenses and emissions profiles often compare favorably. Whether fuel cell technology can scale to gigawatt-level deployment, rather than the hundreds of megawatts installed so far, remains an open question.
What it means for investors
For Brookfield, the expanded deal reinforces AI infrastructure as a near-term growth focus, though it also adds to a complex financial picture that includes thin net margins and a richly valued stock, according to analysis from Simply Wall St. Estimates of Brookfield’s fair value have varied widely among market analysts, underscoring how differently investors are interpreting the company’s expanding bets on AI-related infrastructure.
Whether this model, pairing large scale capital commitments with onsite power generation, becomes the standard approach for AI infrastructure buildout, or simply one option among several, remains to be seen as developers across the industry race to secure power however they can.