Bloom Energy SOFC for AI Data Centers, $5 B Brookfield Deal, 400 MW Deployed, and Oracle Agreements (2021 to 2025)

Data Center On-Site Power Adoption, Bloom Energy SOFC Deals

The urgent need for scalable AI infrastructure has driven data center operators to adopt on-site power solutions to bypass multi-year grid interconnection delays, shifting technologies like solid oxide fuel cells (SOFCs) from backup to primary power sources. Before 2024, fuel cells were primarily considered for supplemental or backup power in commercial and industrial settings. However, the exponential growth in power demand from AI, projected to add 35 GW of demand by 2030, combined with aging grid infrastructure, has created a structural shift. Companies now procure on-site power as a primary, baseload solution to ensure operational continuity and accelerate time-to-market for new data center capacity.

Shift from Grid-Tied to On-Site Primary Power

The fundamental change in 2025 is the use of AI data center fuel cells as “always-on” primary power, a strategy designed to circumvent grid-related development bottlenecks. In the 2021-2024 period, data center power strategies were largely dependent on securing long-term power purchase agreements with utilities and connecting to the existing transmission grid. This model has become untenable, with wait times for new high-voltage connections extending to several years in high-demand regions. In response, major operators are now deploying on-site generation to achieve power independence and operational speed. This pivot is validated by reports that Bloom Energy can deliver and operationalize its systems within 90 days, a significant acceleration compared to traditional grid projects.

Key Commercial Agreements Validating Fuel Cells

Large-scale commercial agreements in 2025 confirm the role of fuel cells as a critical enabler for AI expansion. In July 2025, Oracle announced it would deploy Bloom Energy fuel cells for its Oracle Cloud Infrastructure, specifically to support large-scale AI workloads. Shortly after, data center REIT Equinix announced collaborations with multiple alternative energy providers, including fuel cell developers, to secure its power supply chain. These agreements demonstrate that hyperscalers and colocation providers view on-site power not just as a technical solution but as a core component of their competitive strategy to capture AI market share.

AI Workloads Fuel Exponential Data Center Energy Demand
AI workloads are set to drive data center electricity consumption to nearly 600 TWh by 2028 in a high-growth scenario. Servers – AI Specialized are projected to be the dominant energy consumer, highlighting a critical power demand challenge.

Massive AI Power Demand Strains Grid, Creates On-Site Generation Opportunity
The projected surge in AI-driven data center energy demand will strain existing grid infrastructure, necessitating innovative, efficient, and sustainable on-site power solutions. Hyperscale and large colocation data centers will account for over 90% of server energy consumption by 2028.

(Source: Shehabi et al. — via Equity Macro Strategy: 2026 – by Ridwan Adetu – MetaMacro)

$5 B Brookfield-Bloom Energy AI Infrastructure Partnership

Significant institutional capital is flowing into the on-site power sector, validated by a landmark $5 billion strategic partnership and the creation of specialized financial products designed to fund the AI-related energy build-out. This investment activity signals strong market confidence in the long-term viability of fuel cells as essential infrastructure for the digital economy. The financial backing provides fuel cell manufacturers with the capital needed to scale production and meet the accelerating demand from the data center industry.

The Brookfield $5 B Strategic Partnership

The partnership between Brookfield and Bloom Energy, announced in October 2025, is a pivotal event that structures financing for on-site fuel cell deployments at AI data centers. Under the agreement, Brookfield will fund the distributed power generation assets, which are then leased to customers through a power purchase agreement with Bloom Energy. This financial model de-risks the adoption for data center operators by shifting the upfront capital expenditure to a predictable operational expense, thereby lowering the barrier to entry and accelerating procurement cycles for mission-critical power infrastructure.

Market Validation Through Public Markets

Public markets have responded strongly to the central role of on-site power in the AI value chain. Following the Brookfield deal, Bloom Energy‘s stock experienced a significant increase, reflecting investor recognition of its strategic position. Further validation came in July 2025 with the launch of the Defiance AIPO ETF, the first exchange-traded fund focused specifically on AI and Power Infrastructure. The creation of such a targeted investment vehicle underscores the maturation of this trend from a niche industrial concern into a major, investable market theme.

Table: Key Investments in AI Power Infrastructure (2025)

Bloom Energy Data Center Agreements with Oracle & AEP (2025)

Strategic alliances between fuel cell manufacturers, cloud providers, and utilities in 2025 are accelerating the deployment of on-site power and establishing a new project delivery model for the data center industry. These partnerships are not merely transactional; they represent a coordinated effort to solve a systemic infrastructure challenge. By collaborating directly, these entities are creating a more resilient and responsive power ecosystem tailored to the specific demands of AI workloads.

Oracle Cloud’s Rapid Deployment Strategy

The collaboration between Oracle and Bloom Energy, announced in July 2025, exemplifies the strategic need for speed in the AI sector. Oracle is deploying Bloom‘s SOFC technology to provide clean, reliable power for its cloud infrastructure, explicitly citing the need to support large-scale AI applications. A critical factor in the decision was the ability to deploy the on-site power systems rapidly, enabling Oracle to scale its capacity much faster than if it relied solely on traditional grid connections. By mid-2025, Bloom Energy had already deployed over 400 MW of capacity at data centers globally, demonstrating its execution capability.

AEP’s Utility-as-a-Service Model

In a significant shift in utility strategy, American Electric Power (AEP) received approval in June 2025 to install on-site fuel cells for its large data center customers, including Amazon Web Services (AWS) and Cologix. AEP placed an initial order for 100 MW of fuel cells from Bloom Energy to provide reliable power directly at customer facilities. This move represents a “utility-as-a-service” model, where the utility procures and manages distributed generation assets on behalf of its clients. This strategy allows AEP to retain major industrial customers and manage grid load, while providing data center operators with a streamlined path to secure power.

Table: Key Fuel Cell Partnerships for AI Data Centers (2025)

US Market Focus, Quanta Services Infrastructure Role

The United States is the primary market for the adoption of fuel cells for AI data centers, driven by the high concentration of AI development in the country and significant grid constraints in key regions. While fuel cell technology is global, the most significant commercial deployments for data centers in 2025 are centered in the U.S. This regional focus is creating a substantial market for infrastructure builders like Quanta Services, which are contracted to build the supporting energy and communications systems.

Concentration in US Data Center Hubs

Activity is concentrated in established and emerging U.S. data center hubs where power availability is a primary constraint. The partnerships involving Oracle, AEP, AWS, and Cologix are located in the U.S. This geographic concentration is a direct result of the AI supply chain, which is heavily weighted toward U.S.-based cloud service providers. As these companies build out their AI capacity, they are simultaneously creating localized demand for on-site power solutions and the related construction and engineering services provided by firms like Quanta Services.

Infrastructure Enablement by Quanta Services

While technology firms like Bloom Energy provide the fuel cells, infrastructure solutions providers like Quanta Services play a distinct but equally critical role. Quanta Services does not manufacture fuel cells; it builds the complex infrastructure that supports new power generation and grid modernization. In October 2025, Quanta announced a major agreement with utility Ni Source to provide power generation and grid infrastructure solutions for a large industrial customer. The company’s strong financial results, with revenues of $6.8 billion in Q 2 2025, reflect the high demand for its services in building out the energy ecosystem required to support the AI sector.

SOFC Technology Reaches Commercial Scale for Data Centers

Solid oxide fuel cell (SOFC) technology has matured from niche applications to a commercially viable primary power solution for large-scale data centers, validated by gigawatt-scale production plans and deployments by major cloud providers in 2025. This marks a critical inflection point for the technology. In the 2021-2023 period, fuel cell deployments were often smaller and focused on commercial buildings or as backup power. Today, they are being procured for baseload, mission-critical operations at a scale of hundreds of megawatts.

Validation through Large-Scale Deployments

The technology’s maturity is confirmed by its performance and reliability in demanding data center environments. SOFCs from Bloom Energy are marketed as offering 99.999% uptime, a key requirement for always-on AI services. They also provide high electrical efficiency of 60-65%, significantly better than traditional combustion-based power. These performance characteristics, combined with the successful deployments for clients like Oracle, have validated SOFCs as a bankable and technically sound solution for the data center industry’s power challenges.

Production Capacity Expansion

In response to surging demand, fuel cell manufacturers are aggressively scaling their production capacity. Bloom Energy announced it is on track to double its annual manufacturing capacity to 2 GW by the end of 2026. This expansion is a direct reaction to the growing pipeline of projects from the data center sector and is a strong indicator that the industry expects the trend of on-site power adoption to continue. Scaling production is essential to reducing costs through economies of scale and meeting the delivery timelines required by hyperscale customers.

SWOT Analysis for Fuel Cell Data Center Adoption

While fuel cells for data centers offer a strong solution to grid delays, their growth depends on managing fuel supply risks and scaling manufacturing, with the market opportunity defined by the persistent growth in AI power demand. The technology’s primary strength is its ability to provide rapid, reliable power, directly addressing the main bottleneck for data center expansion. However, its current reliance on natural gas presents a challenge to corporate decarbonization goals, a weakness that the industry is addressing through a transition to hydrogen fuel cells.

Strengths and Opportunities

The key strengths of fuel cells are their rapid deployment capability, high reliability, and modular scalability, which perfectly align with the needs of the fast-growing AI industry. The main opportunity lies in the integration with green hydrogen. As production of green hydrogen scales and costs decline, fuel cells offer a clear pathway to providing zero-carbon, 24/7 power for data centers, positioning the technology as a long-term sustainable solution.

Weaknesses and Threats

The primary weakness is the current dependence on the natural gas pipeline, which exposes operators to price volatility and presents a carbon footprint. The main threat comes from two areas: the potential for competing technologies, such as advanced small modular reactors (SMRs) or long-duration energy storage, to mature and become cost-competitive, and the possibility that large-scale grid modernization and transmission build-outs eventually alleviate the connection backlogs, reducing the urgency for on-site generation.

Table: SWOT Analysis for Fuel Cell Adoption in AI Data Centers

Scenario: Fuel Cell Market Growth with Bloom Energy

The most critical factor for 2026 is whether fuel cell manufacturers can execute on their production scaling plans to meet the exponential demand from data center operators. Any failure to meet delivery timelines for multi-hundred-megawatt orders could create significant openings for alternative on-site power technologies and potentially slow the pace of AI infrastructure deployment.

The High-Growth Scenario

If demand from data centers continues to outpace grid capacity, watch for continued announcements of large-scale, multi-year procurement agreements between fuel cell providers and cloud companies. Key signals to monitor include Bloom Energy successfully meeting its target of 2 GW annual production capacity by the end of 2026 and the announcement of the first major data center project designed to run primarily on green hydrogen-powered fuel cells. This scenario would cement fuel cells as essential data center power for the foreseeable future.

The Constrained-Growth Scenario

If manufacturers face supply chain disruptions or technical challenges in scaling production, this could lead to project delays and customer frustration. In this scenario, watch for data center operators diversifying their on-site power strategies more aggressively, with increased investment in natural gas turbines as a stop-gap measure or accelerated pilot projects for emerging technologies like SMRs. A slowdown in deployment would signal that the fuel cell industry is struggling to capitalize on its market opportunity, potentially ceding ground to competitors.

The questions your competitors are already asking

This report covers one angle of fuel cell adoption for data centers. The questions that matter most depend on your work.

• Data center projects using green hydrogen fuel cells
• Small modular reactors for data center power viability
• FuelCell Energy data center deals
• Cost of on-site power purchase agreements for data centers

This report does not answer these. Enki Brief Pro does.

Your question, your angle, your framework. SWOT, PESTL, scenario modelling. The same niche depth, built around the decision your work actually depends on.

Run your first brief in Enki Brief Pro