Most investors understand that AI needs chips.
More are beginning to understand that AI needs networking.
The next realization is that AI needs power.
Lots of it.
The biggest constraint on AI over the next decade may not be GPUs. It may be electricity.
That is where fuel cells enter the story.
The Project Jupiter Shock
Project Jupiter is not an ordinary data center project.
Located in Doña Ana County, New Mexico, it is part of the broader Stargate initiative involving Oracle, OpenAI, SoftBank, and other strategic partners.
The full campus is expected to support approximately 800,000 GPUs across four hyperscale facilities.
Estimated investment: $165 billion.
Power requirement: Approximately 2.45 gigawatts.
The original design relied on:
Natural gas turbines for primary power
Diesel generators for backup power
In April 2026, Oracle abandoned that plan.
The company signed an agreement with Bloom Energy to supply up to 2.8 GW of fuel-cell capacity, including 2.45 GW for Project Jupiter.
A single AI campus created demand larger than the entire U.S. fuel-cell market was only a few years ago.
Finish the survey, get $60 off Annual Pro Membership.
Why AI Needs More Power Than Ever
Modern AI workloads consume enormous amounts of electricity.
Training frontier AI models requires thousands of GPUs operating continuously for weeks or months. Inference workloads run around the clock, serving millions of users and applications.
The result is a surge in power demand unlike anything the data center industry has experienced before.
Major technology companies are now competing not only for chips but also for access to electricity.
In many regions, utility infrastructure cannot be expanded quickly enough to meet demand.
New transmission lines, substations, and generation facilities often require years to build.
This creates a significant opportunity for alternative power solutions.
Power has become the next major AI bottleneck.
What Are Fuel Cells?
Fuel cells generate electricity through an electrochemical process rather than combustion.
Instead of burning fuel, they convert hydrogen or natural gas into electricity with high efficiency and low emissions.
Continuous 24/7 operation
High reliability
Modular deployment
Lower emissions than traditional generators
Reduced dependence on grid infrastructure
Unlike solar and wind, fuel cells can operate regardless of weather conditions.
Unlike diesel backup generators, they can provide primary power continuously.
Why Fuel Cells Are Suddenly Relevant
Fuel cells are not a new technology.
What is new is the problem they solve.
Instead of waiting years for grid upgrades, operators can deploy on-site power generation in months.
That capability is becoming increasingly valuable as hyperscalers race to build AI capacity.
Fuel cells offer:
• Continuous 24/7 power generation
• High reliability
• Modular deployment
• Reduced dependence on local grid constraints
• Lower emissions than diesel-based alternatives
The technology itself is less important than the outcome.
Fuel cells allow data centers to accelerate deployment.
In an industry where time-to-power increasingly determines time-to-revenue, that advantage matters.
The Three Fuel Cell Stocks Powering the AI
The fuel-cell opportunity is not a single-company story.
Each company occupies a different position within the ecosystem.
Each offers a different risk-reward profile.
→ Get the Full AI Energy Watchlist
This list is shared exclusively with my subscribers.
Bloom Energy ($BE)
The AI Infrastructure Leader
What They Do
Bloom Energy develops solid oxide fuel cells (SOFCs) that generate electricity directly at customer sites.
Their systems can operate on natural gas today and transition toward hydrogen over time.
Bloom has become the clear AI infrastructure leader within the fuel cell industry.
Why Bloom Is Different
Bloom’s value proposition is simple:
Instead of waiting years for grid upgrades, data center developers can deploy Bloom fuel cells in months.
This solves one of the largest bottlenecks facing AI infrastructure.
Financial Performance
Bloom reported:
2025 revenue of $2.02 billion
Revenue growth of 37.3% year-over-year
Gross margin of 29%
Record annual revenue performance
In 2026, momentum accelerated further.
Q1 revenue more than doubled to $751 million and the company raised full-year revenue guidance to $3.4-$3.8 billion. The company also turned significantly profitable as AI-related demand increased.
Strategic Partnerships
Oracle
Expanded partnership supporting up to 2.8 GW of fuel-cell deployment across Oracle's AI infrastructure footprint.
Brookfield Asset Management
Up to $5 billion partnership focused on deploying Bloom systems across AI data-center projects globally.
Investment Thesis
Bloom is increasingly being valued less as a clean-energy company and more as an AI infrastructure company.
If power remains a constraint on AI expansion, Bloom could become one of the most important beneficiaries of the next phase of the AI buildout.
$FCEL - FuelCell Energy
The Asymmetric Challenger
What They Do
FuelCell Energy develops megawatt-scale fuel-cell systems serving utilities, industrial customers, and distributed power markets.
Its platform also supports carbon capture and hydrogen production applications.
Financial Performance
Q1 2026:
Revenue: $30.5 million
Revenue Growth: 61%
Operating losses improved year-over-year
Profitability remains a work in progress
Management increasingly highlighted data-center power demand as a long-term growth opportunity throughout 2025.
AI Opportunity
FuelCell Energy offers exposure to the same secular trend as Bloom but at a much earlier stage.
If AI data center demand creates a massive distributed power market, FuelCell Energy could participate meaningfully.
Unlike Bloom, however, FuelCell Energy has not yet demonstrated comparable commercial traction in AI deployments.
Investment Thesis
FuelCell Energy represents the highest-risk, highest-upside fuel-cell opportunity.
Investors are effectively betting that AI power demand expands faster than the market currently expects and that FuelCell can capture a meaningful share of that growth.
Plug Power ($PLUG)
The Hydrogen Infrastructure Bet
What They Do
Plug Power has built one of the industry's most comprehensive hydrogen ecosystems.
Operations include:
Hydrogen production
Fuel cells
Electrolyzers
Hydrogen logistics
Refueling infrastructure
The company is often viewed as a hydrogen economy investment rather than a pure fuel cell investment.
Key Partnerships
Amazon: Hydrogen-powered warehouse operations
Walmart: Material handling fuel cell deployments
Airbus: Hydrogen aviation initiatives
AI Relevance
Unlike Bloom, Plug is not currently a direct AI infrastructure story.
Its AI exposure depends on a future scenario where:
Hydrogen adoption accelerates
Fuel cells become a larger component of power infrastructure
Data centers increasingly utilize hydrogen-based energy systems
Investment Thesis
Plug is a long-duration bet on hydrogen rather than a direct AI power play.
If hydrogen becomes a major component of global energy infrastructure, Plug’s integrated platform could become extremely valuable.
If adoption remains slow, profitability and financing challenges will likely remain headwinds.
Investors should view Plug as a speculative hydrogen infrastructure play with significant upside potential and elevated execution risk.
The Bigger Opportunity
Fuel cells are probably not the final answer.
They’re simply the first answer.
The larger lesson involves bottlenecks.
Every major technological revolution creates them.
The companies solving bottlenecks often outperform the companies creating demand.
During the AI compute boom, investors focused on Nvidia.
The bottleneck winners included photonics, networking, and memory.
Today the bottleneck is power.
Tomorrow it may be transformers.
Next year it may be nuclear.
The technology changes.
The pattern doesn’t.
Final Thought
Project Jupiter revealed something important.
The largest technology companies in the world are no longer assuming the grid will solve their problems.
They’re building power solutions directly into their AI infrastructure.
That’s a major shift.
Bloom Energy is the current leader.
FuelCell Energy is the asymmetric challenger.
Plug remains a speculative hydrogen bet.
But the bigger takeaway is this:
The first AI bottleneck was compute.
The next AI bottleneck is power.
And capital tends to flow toward bottlenecks long before consensus notices.
The infographic below summarizes the fuel-cell landscape, highlights the
major public companies, and shows how fuel cells are becoming an increasingly important part of the AI infrastructure buildout.
Finish the survey, get $60 off Annual Pro Membership.
Disclosure: For informational purposes only. Not investment advice. It reflects my personal opinions for research and discussion purposes only. I may hold positions mentioned and may change positions at any time without notice. Do your own research.