2026-08-05T18:30:00.000Z
A $2.65 Billion Utility Contract Just Redefined What Hydrogen Fuel Cell Demand Means
American Electric Power disclosed on January 8, 2026, that it had exercised most of its option on Bloom Energy's solid oxide fuel cells, a deal worth about $2.65 billion. Not one unit will move a vehicle. Every megawatt is going to a fuel cell power facility near Cheyenne, Wyoming, built for data center customers that need power faster than the grid can deliver it.
That single contract says more about where this market is headed than any transportation forecast. Kaiso Research's primary dataset puts the global hydrogen fuel cells market at $7.75 billion in 2025, climbing to $62.43 billion by 2035 at a 23.2% CAGR. The report's own segmentation still names transportation and fuel cell vehicles as the largest categories today. The capital moving fastest right now, Oracle's 2.8 gigawatt commitment to Bloom Energy and Brookfield's $25 billion financing framework, is stationary.
For a VP of infrastructure weighing an onsite power decision this quarter, that gap between the stated segment leader and the money's actual direction is the whole story. This article covers both halves: the transportation and defense buildout Kaiso Research tracked through 2025, and the stationary power shift that is rewriting the buyer list in 2026.
Kaiso Research Puts the 2025 Base at $7.75 Billion. Three Forces Explain the Jump to $62.43 Billion
Kaiso Research's primary dataset across the type, application, and end user segments it tracks puts the 2025 hydrogen fuel cells market at $7.75 billion, rising to $62.43 billion by 2035, a 23.2% compound annual growth rate across the 2026 to 2035 forecast period. Three structural forces explain a curve this steep.
The first is technology segmentation. Proton exchange membrane (PEM) fuel cells lead adoption on power density and fast response, which is why Kaiso Research's report names them the dominant type. Solid oxide fuel cells (SOFC) trail in unit count but lead in the megawatt scale that utilities and hyperscalers now buy in single contracts.
The second is application mix. Kaiso Research segments the market into stationary, transportation, and portable use, and its report names transportation as the largest application today, driven by fuel cell electric vehicle (FCEV) programs in Japan, South Korea, and China. The third is geography. Asia-Pacific leads regional share in the Kaiso Research dataset, a position built on Japan's Ene-Farm residential fuel cell installations and South Korea's Hydrogen Economy Roadmap targets for fuel cell vehicles.
None of those three forces required an AI data center to exist. What changed the slope of the curve is a fourth force layered on top: stationary fuel cells being bought as bridge power for electricity demand the grid cannot yet serve.
Bloom Energy Is Building AI Power Plants. Ballard and Plug Power Are Still Fighting for Transportation Share
Ten companies anchor the competitive set in Kaiso Research's report: SFC Energy AG, Ceres, Doosan Fuel Cell, AFC Energy, Nedstack Fuel Cell Technology, Intelligent Energy, Ballard Power Systems, FuelCell Energy, Bloom Energy, and Plug Power. Their 2026 positions have diverged sharply from where the report's company profiles left them.
Bloom Energy sells solid oxide systems, and its 2026 has been defined by data center contracts rather than the commercial and industrial deployments the Kaiso report describes. American Electric Power's original 2024 agreement for 100 megawatts of Bloom's fuel cells carried an option for 900 megawatts more, and AEP exercised most of that option on January 8, 2026, in a deal worth about $2.65 billion for a facility near Cheyenne, Wyoming. On April 13, 2026, Oracle expanded its own commitment to up to 2.8 gigawatts, with 1.2 gigawatts already contracted after Bloom delivered Oracle's first system in 55 days.
Ballard Power Systems occupies the opposite end of the technology spectrum. Its PEM engines power buses, trains, and the mining truck Adani Enterprises deployed in Chhattisgarh in May 2025. Ballard's first quarter 2026 revenue reached $19.4 million, up 26% year over year, with a third consecutive quarter of positive gross margin and $516.8 million in cash. Marty Neese, Ballard's CEO, is steering the company toward bus and rail contracts, not data centers.
Plug Power sits between the two. Its electrolyzer and green hydrogen production business reported $700 million in 2025 revenue and a positive fourth quarter gross margin, a real recovery under CEO Jose Luis Crespo after years of cash concerns that once raised open questions about the company's survival.
Ceres Power and Doosan Fuel Cell occupy a smaller, licensing-driven corner of the same solid oxide technology. Ceres licenses its stack design rather than manufacturing at scale itself, with DENSO Corporation and Thermax signing on as 2024 manufacturing partners. Doosan Fuel Cell supplies phosphoric acid units for Korean power generation and, through a license from Ceres, the solid oxide stacks it is now testing for maritime use.
FuelCell Energy is the newest entrant to the data center trend. It announced a deal with Fit Energy in June 2026 for up to 380 megawatts of on-site power, evidence that Bloom Energy will not be the only supplier this buyer category eventually funds.
The report's company list has not changed. What each company sells its own investors on has changed considerably.
Grid Interconnection Queues Are Now Longer Than AI Buildout Timelines, and Fuel Cells Are the Fix
Multi-year grid interconnection queues, not vehicle subsidies, are the single largest driver behind the acceleration in Kaiso Research's forecast. Data center operators cannot wait three to five years for a new substation, and fuel cell systems can be commissioned in months.
Bloom Energy delivered Oracle's first onsite system in 55 days, more than a month ahead of its own 90 day target. Brookfield, the infrastructure investor managing over $1 trillion in assets, expanded its financing framework with Bloom Energy from $5 billion in October 2025 to $25 billion by June 30, 2026, a fivefold increase built specifically to fund fuel cell deployments that bypass utility timelines entirely. That's not a pilot budget. It is a bet that on-site generation becomes the default, not the exception, for new AI campuses.
Transportation and heavy-duty adoption remain real drivers in their own right. Fuel cell electric vehicles benefit from longer range and faster refueling than battery alternatives in mining, maritime, and long-haul freight applications, and defense programs continue to fund lightweight fuel cells for drones and field power. But the arithmetic has shifted. Gartner's forecast points to 27% growth in worldwide data center power demand in 2026 alone, reaching 132 gigawatts, a single year of digital infrastructure growth that dwarfs a decade of FCEV sales guidance from any automaker covered in this report.
Policy support still matters, but it is regionally uneven. Asia-Pacific government programs continue to underwrite transportation adoption. American and European stationary demand is now being pulled by data center economics first and by climate policy second, a reversal from how this market operated as recently as 2023.
Defense and maritime uses reinforce the case for hydrogen fuel cells beyond AI power and vehicles. The Department of Defense continues to fund lightweight fuel cells for drones and field power, where energy density matters more than refueling speed. Doosan Fuel Cell's maritime SOFC program, tied to Korea Shipbuilding and Offshore Engineering, points to IMO decarbonization targets as a third demand source that no single AI contract can replace.
From a Chhattisgarh Coal Mine to a Wyoming AI Campus: Four Deployments That Actually Shipped
Four 2025 and 2026 deployments show where hydrogen fuel cell adoption is real, not projected, and they correct the dating on developments Kaiso Research's own report describes.
Adani Enterprises flagged off India's first hydrogen fuel cell mining truck in Raipur, Chhattisgarh, on May 10, 2025, not in mid-2024 as the underlying partnership's original announcement date is sometimes conflated with its eventual deployment. The truck, built on an Ashok Leyland platform with a 120 kilowatt Ballard FCmove-XD engine, hauls up to 40 tons of coal across a 200 kilometer route between the Gare Pelma III mine and a state power plant. The original three way agreement between Adani, Ashok Leyland, and Ballard actually dates to January 2023, more than two years before the truck entered service.
Bloom Energy's SK Eternix agreement in South Korea is a second corrected case. The deal, an 80 megawatt installation described as the largest single site solid oxide fuel cell project and the largest project financing for fuel cells in South Korean history, was signed on November 7, 2024, financed by Korea Development Bank, with commercial operations targeted for 2025.
A third deployment sits in the data center pipeline directly. FuelCell Energy signed its 380 megawatt agreement with Fit Energy on June 23, 2026, with an initial 30 megawatts scheduled for delivery within the year, evidence that the AI power trend has spread beyond Bloom Energy to a second listed fuel cell manufacturer. Fourth, Ceres Power's SOEC licensing model gained two manufacturing partners in quick succession: DENSO Corporation in August 2024 and Thermax in India in September 2024, both aimed at scaling solid oxide electrolysis rather than PEM technology.
PEM Wins the Road. Solid Oxide Wins the Data Center. The Difference Is Temperature
Operating temperature, not marketing, decides which fuel cell chemistry wins which application. PEM cells run cool, start fast, and respond well to load changes, which is exactly what a bus, truck, or forklift needs. SOFC units run above 600 degrees Celsius, take longer to reach operating temperature, and reward the buyer with higher electrical efficiency once they are running continuously.
That thermal profile explains the current market split cleanly. Ballard's FCmove engines, the same platform used in the Adani mining truck, are built for PEM's rapid response and compact footprint, ideal for vehicles that start and stop throughout a shift. Bloom Energy's Energy Servers and Ceres Power's SteelCell stacks are SOFC designs built to run at a constant, high output for months at a time, which is precisely the profile a data center or an eco park wants.
Cost curves are moving in opposite directions for the two technologies. PEM stacks still depend on platinum group catalysts, and cutting that cost has been the industry's central research focus for a decade. SOFC systems avoid platinum entirely but historically carried higher balance of plant costs, a gap Bloom Energy is closing faster than PEM catalyst costs are falling by working to double its own manufacturing capacity to 2 gigawatts annually.
Buyers evaluating either technology in 2026 are not choosing based on which is more advanced. They are matching thermal and duty cycle requirements to the chemistry built for them.
Bloom Energy's Contracted Capacity Now Dwarfs Every Transportation Fuel Cell Maker Combined
Bloom Energy's 2.8 gigawatt Oracle commitment and its roughly $2.65 billion AEP agreement alone represent more contracted value than Ballard, Plug Power, and FuelCell Energy have shipped across their entire operating histories combined. That's not a knock on the transportation side of this market. It is a statement about where investor capital has actually gone in 2026.
Ballard holds the strongest position among pure transportation players, backed by $516.8 million in cash, a New Flyer bus contract, and 300 million kilometers of real world fleet operation on its engines. Plug Power's recovery is real but slower, still working through tariff exposure and years of accumulated losses even as 2025 revenue crossed $700 million. Ceres Power has chosen not to compete for manufacturing scale at all, licensing its SOEC technology to DENSO and Thermax instead and collecting royalty economics.
FuelCell Energy is the company to watch for a possible repeat of Bloom Energy's trajectory. Its Fit Energy agreement signals the data center opportunity is not exclusive to one supplier, and any second or third entrant that proves it can deliver gigawatt scale power on a data center's timeline will attract the same capital Bloom Energy has. The transportation fuel cell makers are not losing this market. They are simply not the ones setting its growth rate anymore.
Brookfield's $25 Billion Bet Dwarfs the Entire DOE Hydrogen Hub Program
Private capital committed to hydrogen fuel cell projects in the first half of 2026 has already outpaced a decade of announced U.S. federal hydrogen funding. Brookfield's financing framework with Bloom Energy, expanded from $5 billion to $25 billion on June 30, 2026, exists to fund fuel cell projects for AI infrastructure specifically, not the broader clean energy mandate that shapes public funding decisions.
Compare that figure to the U.S. Department of Energy's Regional Clean Hydrogen Hubs program, authorized for up to $7 billion across seven hubs in October 2023. Phase 1 awards disbursed since then have run far below that headline number, with individual hubs receiving between $18.8 million and $22.2 million for planning and design work as of early 2025, and later phases remain subject to renegotiation. The European Hydrogen Bank's third auction, by contrast, closed in February 2026 having awarded just over €1 billion to nine projects, a result 6 to 7 times oversubscribed against its available budget.
Public hydrogen funding is not disappearing. It is simply being outpaced by private capital that has found a faster paying customer in AI infrastructure than in either transportation subsidies or industrial decarbonization mandates.
The One Big Beautiful Bill Act Cut the 45V Credit Window by Five Years. Europe Kept Bidding Anyway
Legislative timelines for hydrogen incentives now differ sharply between the United States and Europe, and that asymmetry is a decision input, not background noise. In the United States, the One Big Beautiful Bill Act compressed the Section 45V clean hydrogen production credit's construction deadline from January 1, 2033 to January 1, 2028, forcing developers to accelerate or abandon projects. A companion 30% investment tax credit for fuel cell property under Section 48E remains available, but only for systems meeting domestic content thresholds rising from 45% in 2026 to 55% by 2028.
The Department of Energy's Regional Clean Hydrogen Hubs program faces separate uncertainty. Executive actions tied to federal spending reviews paused some disbursements in early 2025, later narrowed by a federal court, and DOE ordered hub recipients to halt Community Benefits Plan work that same January. None of the seven hubs has been cancelled, but timelines have slipped.
Europe's approach has stayed comparatively stable. The European Commission's third Hydrogen Bank auction closed in February 2026, awarding over €1 billion to projects across seven countries and drawing bids worth six times that budget. Brussels is still funding hydrogen production at scale while Washington narrows its own credit window.
Utilities Should Move Now. Automakers Should Wait for Someone Else to Build the Stations
Two buyer groups need opposite strategies from this data, and treating them identically is the most common mistake a strategy team can make reading this market.
Utilities, data center developers, and colocation providers facing multi-year interconnection queues should treat fuel cells as available now, not as a future technology to evaluate. Bloom Energy's 55 day deployment record and Brookfield's willingness to finance projects at $25 billion scale mean the financing and delivery risk that stalled stationary fuel cells for a decade has largely been solved by someone else's balance sheet. A VP of infrastructure who waits another budget cycle to evaluate this option is choosing to lose the interconnection queue race to a competitor who did not wait.
Automotive OEMs and fleet operators face the opposite calculus. FCEV adoption remains real in Japan, South Korea, and heavy-duty niches like mining and maritime, but Kaiso Research's own transportation dominant segmentation reflects committed government programs more than a customer buying signal outside Asia-Pacific. A CFO evaluating fuel cell truck fleets in North America or Europe should wait for refueling infrastructure that other players, not the fleet operator, are incentivized to build first. Ballard's own order book shows why patience pays here: its strongest 2026 wins are transit agencies buying through New Flyer, not private fleets absorbing refueling buildout costs alone.
Three Risks the Kaiso Intelligence Box Cannot Price In
A CAGR is a smoothed line through a market that is not actually smooth, and three specific risks sit underneath the 23.2% figure that a single number cannot show.
The first is trade exposure. Plug Power's fuel cell business depends on components sourced from China, and its electrolyzer imports from Europe are now subject to 20% tariffs, a cost structure that directly offsets the margin gains the company has spent two years building. The second is conversion risk on public funding. Only six of eighteen projects awarded in the European Hydrogen Bank's second auction round actually signed grant agreements, meaning a headline funding award is a weaker signal of real deployment than it appears.
The third is U.S. policy compression. Developers now have until 2028, not 2033, to break ground under Section 45V, and any project that cannot hit that timeline loses the credit entirely rather than receiving a reduced one. Stack those three risks against a market still recording double digit growth, and the honest read is that the CAGR is directionally right while the path to it will be considerably rockier for individual companies than the smoothed number implies.
Where the $62.43 Billion Actually Lands by 2035
Kaiso Research's forecast puts the hydrogen fuel cells market on a path from $7.75 billion in 2025 to $62.43 billion in 2035, and the more useful read of that decade is which half of the market gets there first. Stationary power, led by SOFC contracts tied to AI infrastructure, should compound faster through the back half of this decade than the forecast's average CAGR implies, pulled by Brookfield and Oracle scale commitments that already exceed $25 billion combined.
Transportation and portable applications, concentrated in Asia-Pacific FCEV programs and heavy-duty niches, will keep growing on government roadmaps and industrial decarbonization mandates, just on a flatter curve than stationary power's. By 2035, expect the market's center of gravity to have shifted further toward electricity generation and further away from the vehicle story that built this industry's early reputation.
The Market Kaiso Research Measured in 2025 Is Not the Market That Will Exist in 2030
This report exists because Kaiso Research measured a market defined by fuel cell electric vehicles, government hydrogen roadmaps, and industrial decarbonization pilots. That market still exists. It still matters to the automakers, defense contractors, and Asia-Pacific policymakers building on it, and none of them should abandon it because one utility signed a large contract. It is simply no longer the part of the story moving the fastest.
The companies that understood this fastest already have the contracts to prove it. Bloom Energy didn't out-compete Ballard or Plug Power at their own game. It found a buyer, AI infrastructure operators locked out by grid interconnection queues, that the original hydrogen fuel cell market never priced in. Every fuel cell maker still competing purely for transportation share in 2026 is competing for a slower growing piece of a market whose fastest growing piece now belongs to someone else, and the Kaiso Intelligence Box at the top of this report will need updating again before 2027.
---
About Kaiso Research and Consulting
Kaiso Research and Consulting is a global market intelligence firm publishing 5,000+ research reports across 11+ industry verticals.
[email protected] | +1 872 219 0417
Dhwani Sharma, Lead Industry Analyst, Kaiso Research and Consulting | Covering clean energy and industrial technology markets across North America, Europe, and Asia-Pacific
Published: 2026-07-04 | Report Code: EPGA1359
Market Study: Access the full index or request a complimentary sample directly via the Hydrogen Fuel Cells Market Size, Growth Industry Report, 2026 - 2035 page
