Brookfield Asset Management announced Tuesday that it has expanded its financing partnership with Bloom Energy from $5 billion to $25 billion, with the goal of accelerating on-site power generation for hyperscalers and AI data center developers facing lengthy grid interconnection queues.
From Backup Power to Primary Infrastructure
The two companies first established their financing framework in October 2025. The fivefold expansion is designed to fund Bloom fuel-cell deployments globally while compressing project timelines by pairing capital directly with on-site generation. Rather than waiting for utility connections that can take years to secure, AI campus operators can now bring power online from day one alongside compute and data center infrastructure.
On-site generation has traditionally served as a backup or bridging solution. This partnership treats it as primary infrastructure, a shift that reflects how severely grid delays now constrain AI development schedules.
Energy Certainty as a Financeable Asset
Neil Osnato, founder of Persistence Analytics Group, told Data Center Knowledge that the expansion represents something more nuanced than a new standalone asset class. “I don’t think behind-the-meter power should automatically be viewed as a new asset class in isolation,” he said. “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.”
Brookfield frames the deal within its AI Infrastructure Fund, launched in late 2025 with a stated deployment target of $100 billion across AI factories, power solutions, compute infrastructure, and strategic capital partnerships. The company says it has already committed more than $100 billion to digital infrastructure and clean power assets.
A Competitive Market for Dedicated AI Power
Brookfield and Bloom are not alone in pursuing dedicated power infrastructure for large AI loads. GE Vernova’s gas turbine work for AI campuses and Wärtsilä’s large-scale off-grid data center project in Texas reflect broad industry interest in bringing generation closer to compute workloads. The Brookfield-Bloom model aims to differentiate by embedding financing into the deployment package, making capital a tool for schedule certainty rather than simply a way to purchase equipment.
Risks Still to Be Validated
Osnato cautioned that the next phase of this investment model will require scrutiny of several open questions: whether projected AI loads actually materialize, whether fuel supply and operating costs remain sustainable, how on-site systems interact with long-term utility planning, and who bears stranded-asset risk if demand shifts unexpectedly.
For hosting providers and data center operators, the broader takeaway is practical. Access to bundled financing for on-site generation may prove as strategically important as access to the generation technology itself. As interconnection queues lengthen and AI workloads intensify competition for reliable megawatts, developers who can secure both power and capital simultaneously stand to bring capacity online faster than those waiting on utility timelines alone.