AI Power GridIndustry ShiftJul 1, 2026, 7:20 AM· 4 min read· #2 of 2 in business

Bloom Energy and Brookfield Expand AI Power Partnership to $25 Billion to Bypass Grid Bottlenecks

Brookfield Asset Management is quintupling its commitment to Bloom Energy's solid oxide fuel cells, creating a $25 billion framework to provide rapid, off-grid power for global AI data centers.

By Factlen Editorial Team

AI Infrastructure Developers 45%Clean Energy Advocates 30%Utility Operators 25%
AI Infrastructure Developers
View off-grid power as the only viable way to bypass multi-year utility delays and deploy AI compute capacity on schedule.
Clean Energy Advocates
Emphasize the environmental benefits of fuel cells, which operate without combustion and produce virtually zero smog-forming emissions.
Utility Operators
Recognize that 'behind-the-meter' solutions relieve the immense strain that gigawatt-scale AI facilities place on aging public grids.

What's not represented

  • · Local zoning boards managing data center land use
  • · Traditional utility companies losing large-load customers to off-grid solutions

Why this matters

The AI revolution is currently gated by a severe shortage of electricity and multi-year delays in connecting to the traditional power grid. This historic $25 billion investment proves that the tech industry is successfully pivoting to independent, on-site clean energy to keep AI development moving forward without straining local utilities.

Key points

  • Brookfield Asset Management is increasing its financing framework for Bloom Energy power projects from $5 billion to $25 billion.
  • The capital will be used to deploy solid oxide fuel cells at AI data centers globally, providing reliable, on-site electricity.
  • The expansion is driven by severe constraints on the traditional power grid, where new connections can take up to four years.
  • Bloom's fuel cells bypass the grid entirely, allowing AI developers to bring new compute facilities online in a matter of months.
$25 billion
Expanded financing framework
$5 billion
Original October 2025 commitment
$100 billion
Brookfield AI Infrastructure Fund target
12%
Bloom Energy post-market stock jump
24 to 48 months
Typical grid interconnection delay

Brookfield Asset Management and Bloom Energy have dramatically expanded their strategic partnership, quintupling a previous financing framework to a staggering $25 billion [1][2]. The deal, announced Tuesday, aims to deploy Bloom’s solid oxide fuel cells at artificial intelligence data centers globally, providing a massive injection of capital to solve the tech industry's most pressing bottleneck: electricity.[1][2]

The expansion builds on an initial $5 billion agreement struck in October 2025 [5]. Since then, the sheer scale of AI power demand has outpaced even the most aggressive forecasts. Brookfield’s decision to increase the envelope to $25 billion signals that fully integrated "AI factories" are now a top strategic priority for the world’s largest infrastructure investors, and that securing independent power is the gating factor for their development [3].[3]

Following the announcement, shares of Bloom Energy surged 12% in post-market trading [2]. The market reaction underscores a broader shift in how Wall Street views energy infrastructure. Fuel cells are no longer being treated as niche backup systems; they are rapidly becoming the primary baseload power source for multi-billion-dollar compute facilities [4].[2]

Brookfield is allocating a quarter of its $100 billion AI Infrastructure Fund specifically to Bloom Energy's off-grid power solutions.
Brookfield is allocating a quarter of its $100 billion AI Infrastructure Fund specifically to Bloom Energy's off-grid power solutions.

The core issue driving this $25 billion commitment is the physical limitation of the traditional electrical grid. Connecting a new, gigawatt-scale AI data center to the national transmission system can involve interconnection delays stretching from 24 to 48 months [4]. For hyperscalers racing to train the next generation of AI models, waiting years for a utility substation upgrade is a non-starter.

Bloom Energy’s solid oxide fuel cells (SOFCs) offer a "behind-the-meter" alternative. These systems can be deployed in a matter of months, allowing data center operators to implement a "Bring Your Own Power" strategy [4]. By generating electricity on-site through an electrochemical process rather than combustion, the fuel cells bypass the grid entirely while avoiding the multi-year regulatory rate cases associated with utility power.

Bloom Energy’s solid oxide fuel cells (SOFCs) offer a "behind-the-meter" alternative.

The technology is uniquely suited to the "spiky" nature of AI workloads. A single server rack's power draw can fluctuate wildly during intense compute cycles, creating massive swings that traditional grids struggle to balance. Fuel cells, often paired with supercapacitors, can handle these instantaneous load changes while maintaining the 99.999% uptime required for mission-critical digital infrastructure [4].

On-site fuel cells can be deployed in a fraction of the time it takes to secure a new utility grid connection.
On-site fuel cells can be deployed in a fraction of the time it takes to secure a new utility grid connection.

The $25 billion framework is not a single lump-sum check, but rather a multi-year finance envelope drawn from Brookfield’s dedicated AI Infrastructure Fund. That fund, launched in November 2025, carries a $100 billion deployment target [1][3]. The fact that a quarter of that target is now earmarked specifically for Bloom's power generation highlights that the energy slice of the AI pie is becoming just as critical—and capital-intensive—as the silicon compute slice.[1][3]

"When we formed this partnership, we said it was the first phase of a much larger vision," noted Aman Joshi, Chief Commercial Officer of Bloom Energy [1][5]. The expanded collaboration reflects sustained demand from hyperscalers who require fast, reliable, and community-friendly power that doesn't trigger local zoning battles over massive transmission lines.[1]

Environmental considerations also play a role in the technology's rapid adoption. Because solid oxide fuel cells operate without combustion, they produce virtually zero nitrogen oxide or sulfur oxide emissions. Furthermore, the high heat generated by the fuel cells can be captured and routed into absorption chillers to cool the data center's liquid cooling loops, dramatically improving the facility's overall power usage effectiveness [4].

AI workloads require significantly more power and specialized cooling than traditional cloud computing infrastructure.
AI workloads require significantly more power and specialized cooling than traditional cloud computing infrastructure.

Sikander Rashid, Head of AI Infrastructure at Brookfield, stated that the expansion reflects the conviction behind their broader strategy to deliver fully integrated AI factory solutions [1]. By bundling the real estate, the compute hardware, and the off-grid power generation into a single financed package, Brookfield is rewriting the playbook for how digital infrastructure is built.[1]

The ripple effects of this deal extend beyond Bloom and Brookfield. It validates a new commercial model for the energy transition, proving that private capital is willing to step in and fund distributed, clean power at a scale previously reserved for state-backed utility mega-projects [3]. As AI continues to reshape the global economy, the ability to manufacture reliable electricity on demand has become the ultimate competitive advantage.[3]

How we got here

  1. 2008

    Bloom Energy commercially launches its proprietary solid oxide fuel cell system for distributed power generation.

  2. October 2025

    Brookfield and Bloom announce an initial $5 billion partnership to power AI data centers.

  3. November 2025

    Brookfield officially launches its dedicated AI Infrastructure Fund with a $100 billion deployment target.

  4. June 2026

    The partnership is expanded fivefold to $25 billion to meet surging hyperscaler demand for off-grid power.

Viewpoints in depth

AI Infrastructure Developers

View off-grid power as the only viable way to bypass multi-year utility delays and deploy AI compute capacity on schedule.

For hyperscalers and data center developers, the math is simple: AI chips are useless without power, and the traditional grid is moving too slowly. With interconnection queues stretching up to 48 months in major markets, developers are increasingly willing to pay a premium for 'Bring Your Own Power' solutions. By utilizing on-site fuel cells, they can transform a multi-year utility waiting game into a predictable, months-long equipment deployment, ensuring their multi-billion-dollar AI investments begin generating returns immediately.

Clean Energy Advocates

Emphasize the environmental benefits of fuel cells, which operate without combustion and produce virtually zero smog-forming emissions.

Environmental groups and sustainability officers favor solid oxide fuel cells over traditional diesel or natural gas backup generators because of their clean operating profile. Since the fuel cells generate electricity through an electrochemical reaction rather than burning fuel, they eliminate the nitrogen oxide and sulfur oxide emissions that plague combustion engines. Furthermore, the ability to capture the exhaust heat from the fuel cells and use it to power absorption chillers for data center cooling significantly improves the overall energy efficiency of the facility.

Utility Operators

Recognize that 'behind-the-meter' solutions relieve the immense strain that gigawatt-scale AI facilities place on aging public grids.

While utilities traditionally want to capture large industrial loads, the sheer scale and speed of AI power demand has overwhelmed many regional grids. A single AI data center can demand as much power as a small city, requiring massive upgrades to substations and high-voltage transmission lines. Utility operators quietly welcome behind-the-meter solutions like Bloom's fuel cells because they allow economic development to proceed without forcing the utility to rapidly overhaul its infrastructure or pass the costs of grid upgrades onto residential ratepayers.

What we don't know

  • Whether the global supply chain for solid oxide fuel cell manufacturing can scale fast enough to meet the massive $25 billion demand.
  • How long-term fluctuations in natural gas prices will affect the operating costs of these off-grid AI factories compared to traditional utility rates.

Key terms

Solid Oxide Fuel Cell (SOFC)
A device that produces electricity directly from oxidizing a fuel (like natural gas or hydrogen), offering high efficiency and low emissions without combustion.
Behind-the-meter
Power generation systems installed on the energy consumer's side of the utility meter, allowing them to operate independently of the main public grid.
Hyperscaler
Massive cloud service providers, such as Amazon, Microsoft, and Google, that operate data centers at a global scale.
Baseload power
The minimum, continuous amount of electrical power needed to be supplied to a facility or grid at any given time, requiring 24/7 reliability.

Frequently asked

Why do AI data centers need so much power?

AI training and inference require specialized, energy-intensive chips (like GPUs) that draw significantly more electricity than traditional cloud computing servers, creating massive, concentrated power loads.

Why can't data centers just use the regular power grid?

The sheer volume of power needed, combined with multi-year delays for building new utility substations and transmission lines, makes the traditional grid too slow to keep up with the rapid pace of AI development.

Are solid oxide fuel cells clean energy?

Yes. They generate electricity through an electrochemical reaction rather than combustion, resulting in virtually zero smog-forming emissions (like nitrogen oxides) and operating at very high efficiency.

Sources

Source coverage

3 outlets

3 viewpoints surfaced

AI Infrastructure Developers 45%Clean Energy Advocates 30%Utility Operators 25%
  1. [1]Investing.comAI Infrastructure Developers

    Brookfield and Bloom Energy Expand AI Infrastructure Partnership to $25 Billion

    Read on Investing.com
  2. [2]Seeking AlphaAI Infrastructure Developers

    Bloom Energy, Brookfield expand AI infrastructure partnership by 5x to $25B

    Read on Seeking Alpha
  3. [3]AI Intelligence HubAI Infrastructure Developers

    Where the 25 billion dollars actually goes: Brookfield's bet on Bloom Energy

    Read on AI Intelligence Hub
Stay informed

Every angle. Every day.

Get business stories with full source coverage and perspective breakdowns delivered to your inbox.