The Quiet, Critical Role of Hydrogen in Power Generation

Nicholas Cupps
Apr 30
12 min read

A practical look at hydrogen-cooled generators, data center power demand, utility reliability, and the case for on-site hydrogen production.

When most people talk about hydrogen, they talk about the future.

They talk about fuel cell trucks. Hydrogen cars. Green steel. Ammonia. Long-duration energy storage. Aviation. Shipping. Industrial decarbonization. The hydrogen economy is usually described as something that is still coming, still developing, still waiting for infrastructure, demand, policy support, and lower production costs.

But that misses something important.

Hydrogen is not only a future fuel. Hydrogen is already being used today in some of the most important power assets in the world.

One of the least understood, but most practical, uses of hydrogen is in electric generator cooling.

At large power plants, especially utility-scale plants and large industrial generating assets, hydrogen is commonly used as a cooling gas inside electric generators. This is not because hydrogen is trendy. It is not because of a tax credit. It is not because someone wants to make a marketing claim about clean energy.

It is because hydrogen works.

Hydrogen has physical properties that make it extremely valuable inside large rotating electrical equipment. It has high thermal conductivity, low density, and low viscosity. In simple terms, it removes heat very effectively while creating less drag on the rotor than air. That matters because large generators spin at high speeds, produce enormous amounts of heat, and must operate within tight temperature limits to protect windings, insulation, seals, bearings, and other critical components.

Baker Hughes summarizes the reason clearly: hydrogen is used in large electricity generators because its high thermal conductivity and low viscosity help prevent efficiency losses from drag, or “windage,” on the rotor. Maintaining hydrogen purity is also important for safety, efficiency, and long-term generator reliability. ^[1]

This is a quiet but important part of the power system. And as the world builds more power generation for data centers, AI, industrial electrification, utilities, and large behind-the-meter energy users, this quiet use case for hydrogen deserves much more attention.

Because if the future requires more power, then it also requires more reliable power-generation infrastructure. And if that infrastructure depends on hydrogen-cooled generators, then hydrogen supply is no longer a small operational detail.

It becomes a reliability issue.

The Power Problem Is Getting Bigger

The electricity conversation has changed dramatically in the last few years.

For decades, U.S. electricity demand grew slowly. Utilities planned for gradual load growth. Industrial customers wanted reliability and price stability. Data centers existed, but they were small and did not reshape regional grid planning.

That world is gone.

AI, cloud computing, electrification, reshoring, advanced manufacturing, and massive data center campuses are putting new pressure on the power system.

The Department of Energy reported that U.S. data centers consumed about 4.4% of total U.S. electricity in 2023, and projected that they could consume approximately 6.7% to 12% by 2028. The same DOE summary said data center electricity use increased from 58 TWh in 2014 to 176 TWh in 2023, with a projected increase to 325 to 580 TWh by 2028. ^[2]

The International Energy Agency has also warned that data center electricity demand is accelerating quickly. In its 2026 reporting, the IEA said electricity demand from data centers rose 17% in 2025, while AI-focused data centers grew even faster. In its broader Energy and AI analysis, the IEA projected that global data center electricity consumption could more than double to around 945 TWh by 2030, with the United States accounting for the largest share of the increase. ^[3] ^[4]

This is why some of the largest technology companies and developers are no longer only asking, “Where can we lease space?”

They are asking, “Where can we get power?”

In many markets, power availability has become the constraint. Grid interconnection queues are long. Transmission development is slow. Utilities are under pressure to serve new load while maintaining reliability for existing customers. New generation takes years to permit, finance, build, interconnect, and commission.

That is pushing data center developers toward on-site or near-site power generation.

A 2025 data center power report found that industry leaders expected roughly 30% of all data center sites to use some on-site power as a primary energy source supplemental to the grid by 2030, more than double the expectation from only seven months earlier. The same report noted that announced on-site power capacity had reached 8.7 GW globally, including 4.8 GW expected before 2030. ^[5]

This trend matters for hydrogen.

Many of these new on-site power assets will use large turbines and generators. Some will be gas-fired. Some may eventually blend or burn hydrogen as fuel. Some will be utility-scale plants built to serve dedicated data center campuses. Others will be industrial or behind-the-meter assets designed to provide resilience, reduce exposure to grid congestion, and accelerate time to power.

But regardless of the fuel, all large generators need hydrogen cooling.

Why Hydrogen Is Used in Generator Cooling

Large generators create heat. A lot of heat.

When a turbine spins a generator, the generator converts mechanical energy into electrical energy. Inside that machine, the rotor, stator, windings, insulation systems, and other components must operate within precise thermal limits. Excess heat reduces efficiency, accelerates wear, damages insulation, and shortens equipment life.

Air can cool smaller machines. But as generator size increases, air cooling becomes less attractive. Air is denser than hydrogen, which creates more drag. That drag consumes energy and creates additional heat. In large machines, even small efficiency losses matter.

Hydrogen solves several problems at once.

First, hydrogen transfers heat very effectively. Its thermal conductivity is far higher than air, allowing it to remove heat from critical components more efficiently.

Industry sources commonly describe hydrogen’s cooling performance as roughly 7 to 10 times better than air because of its high thermal conductivity and favorable physical properties. ^[6]

Second, hydrogen is very light. Because it has low density, it reduces windage losses inside the generator. Windage losses are the energy losses caused by the rotor pushing through gas as it spins. In a large generator, reducing drag can improve efficiency and reduce heat generation.

Third, hydrogen helps enable higher output in a smaller machine. A hydrogen-cooled generator can often support higher power density than an air-cooled design. That matters when space, efficiency, capital cost, and operating performance are all important.

Fourth, hydrogen cooling is a proven technology. This is not experimental. Hydrogen-cooled generators have been used in power generation for nearly a century.

Power Engineering has described hydrogen cooling as a safe and effective method for keeping generators cool while minimizing efficiency losses, provided plants follow proper handling, monitoring, and maintenance practices. ^[7]

This last point is important.

The hydrogen discussion often gets trapped in debates about future adoption. But in generator cooling, the use case is already established. Utilities, independent power producers, and large industrial power users are not being asked to take a speculative leap. They are being asked to improve how they source and manage a gas they may already rely on.

The Overlooked Risk: Hydrogen Supply

For a hydrogen-cooled generator, hydrogen purity, pressure, and availability are operationally important.

Hydrogen is not just placed in the generator once and forgotten. Plants must monitor purity. They must manage leakage. They must purge safely during maintenance. They must maintain proper pressure. They must refill or make up hydrogen as needed. They must ensure that oxygen contamination is avoided and that proper safety procedures are followed.

This means the hydrogen supply chain matters.

Historically, many facilities have relied on delivered hydrogen. Depending on the site, that may mean cylinders, tube trailers, bulk delivery, or other industrial gas supply arrangements. For many facilities, this has worked well enough.

But the energy world is changing.

More power assets are being built. More are being operated in mission-critical applications. More customers are demanding uptime. Data centers are not casual power users. They are among the most reliability-sensitive customers in the world. A utility-scale generator serving a data center campus, or a behind-the-meter power plant supporting AI workloads, does not have the same tolerance for operational disruption as a non-critical industrial load. If power is cut off to a data center, even momentarily, the data center shuts down.

For data centers, downtime can be catastrophic. For utilities, forced outages can affect grid reliability, capacity obligations, reserve margins, and customer confidence. For self-generating industrial customers, an outage can shut down production.

So the question becomes simple:

If hydrogen is required to keep the generator operating safely and efficiently, why would a critical power asset rely entirely on a delivered hydrogen supply chain that may be exposed to trucking constraints, weather, driver availability, regional shortages, supplier allocation, transportation cost, or emergency delivery risk?

This is where the on-site model becomes compelling.

The Case for On-Site Hydrogen

On-site hydrogen production changes the supply chain.

Instead of producing hydrogen somewhere else, compressing it, liquefying it, trucking it, storing it, and handling repeated deliveries, the customer produces hydrogen at or near the point of use. For generator cooling, this usually means producing high-purity gaseous hydrogen on-site, then feeding the plant’s hydrogen system as needed.

This does not mean every facility needs a massive hydrogen plant. Generator cooling does not require the same volume of hydrogen as a refinery, ammonia plant, or transportation fueling hub. But it does require reliable, high-purity hydrogen when the plant needs it.

That is exactly the kind of use case where modular, right-sized on-site hydrogen production can make sense.

On-site hydrogen generators are safe, reliable, and cost-effective means to provide hydrogen at the customer’s facility. Nel’s generator-cooling literature describes hydrogen generators producing ultra-high-purity hydrogen gas at 99.999+% purity, with output pressures up to 30 barg and low dew points suitable for generator applications. ^[8] ^[9]

For the customer, the value proposition is not only about the molecule.

It is about control.

An on-site system can reduce dependence on delivered hydrogen logistics. It can reduce the need for frequent cylinder handling or truck delivery coordination. It can provide a dedicated source of hydrogen for a dedicated power asset. It can improve visibility into hydrogen availability. It can be designed around the specific consumption pattern of the generator rather than forcing the customer into the economics and constraints of bulk delivery.

For a utility, that can mean fewer operational headaches.

For a data center power plant, it can mean one less supply chain risk attached to mission-critical uptime.

For a self-generating industrial customer, it can mean more control over a small but essential input to power reliability.

Why This Matters for Data Centers

Data centers are becoming power developers.

That statement would have sounded strange a decade ago. Today, it is increasingly true.

The largest data center operators are no longer simply buying power from the grid and negotiating utility service agreements. They are exploring dedicated substations, long-term power purchase agreements, behind-the-meter generation, fuel cells, gas turbines, nuclear partnerships, storage, microgrids, and hybrid systems.

This is happening because time to power has become a competitive advantage.

If two data center developers are competing for AI workloads, and one can secure hundreds of megawatts of reliable power years earlier than the other, that power advantage becomes a business advantage.

Recent market activity shows how quickly this is moving. NextEra has discussed large gas-fired power projects backed by Japan to support U.S. data center growth, with potential capacity approaching 10 GW. Its Florida Power & Light utility has reported a data center power demand pipeline of 21 GW, with more than half in advanced phases for completion by 2028. Separately, Reuters reported on plans to transform the former Homer City coal plant in Pennsylvania into a natural gas-powered data center campus producing nearly 4.5 GW of power by 2027. ^[10] ^[11]

These are enormous numbers.

When power projects reach this scale, every reliability input matters. Fuel supply matters. Cooling water matters. Turbine maintenance matters. Transformer availability matters. Grid interconnection matters. Spare parts matter.

Hydrogen for generator cooling belongs in that same conversation.

It may not be the largest line item. It may not get the headline. But if a hydrogen-cooled generator cannot maintain proper hydrogen purity and pressure, the plant has a problem.

That makes on-site hydrogen a natural fit for data center power projects.

Not because data centers want to be in the hydrogen business. They do not.

They want power.

They want uptime.

They want speed.

They want fewer bottlenecks.

They want partners who can remove operational complexity.

That is where a company like Davy can position itself: not as another hydrogen hype company, but as a practical infrastructure partner solving a real reliability problem for power-intensive customers.

Why This Matters for Utilities

Utilities are under a different kind of pressure.

They are being asked to serve rapidly growing load while maintaining affordability, reliability, and regulatory credibility. They must plan around data centers, electrification, industrial development, extreme weather, aging infrastructure, renewable integration, and reserve margin concerns.

NERC’s 2025 Long-Term Reliability Assessment warned that resource adequacy risks are intensifying across North America, with uncertainty around new resource additions and rising electricity demand from data centers and large loads creating concerns that the industry may struggle to keep pace. ^[12]

For utilities, the value of on-site hydrogen is not about novelty. It is about reliability, standardization, and operational control.

A utility operating hydrogen-cooled generators already understands the importance of gas purity, pressure, safety procedures, purging protocols, leak detection, and preventive maintenance. The question is whether the hydrogen supply model has evolved to match the reliability expectations placed on the power system.

On-site hydrogen can support utility operations in several ways.

It can provide a dedicated source of high-purity hydrogen for generator make-up gas. It can reduce exposure to supply interruptions. It can limit reliance on emergency deliveries. It can simplify logistics at remote plants. It can be incorporated into planned maintenance and outage procedures. It can reduce the need to store large quantities of delivered gas. And if designed properly, it can become part of a broader plant reliability strategy.

This is especially relevant for peaking plants, combined-cycle plants, and critical generation assets serving constrained regions.

As utilities add new gas-fired generation to support data centers and grid reliability, some of those assets may be hydrogen-cooled. Even where the generator cooling demand is modest compared with the plant’s fuel demand, the consequences of hydrogen unavailability can be significant.

In other words, the volume may be small.

The importance may be large.

Why This Matters for Self-Generating Customers

There is also a third customer segment: large users generating their own power.

These may include industrial facilities, campuses, manufacturing plants, hospitals, universities, military installations, and large commercial or technology sites. Increasingly, some of these customers are exploring on-site generation because they cannot wait for the grid to catch up, or because they need a higher level of resilience than the grid alone can provide.

For these customers, power reliability is business continuity.

If the plant goes down, production stops. If the campus loses power, operations are disrupted. If backup systems fail, the consequences can be expensive or dangerous.

A self-generating customer may not want to manage hydrogen supply. They may not have internal expertise in industrial gases. They may not want cylinders, delivery schedules, emergency orders, supplier calls, or uncertainty about purity and availability.

They want a turnkey solution.

This is where Davy’s model can be especially powerful.

A modular on-site hydrogen system can be built around the customer’s actual use case.

It can be sized for the generator cooling requirement. It can include storage, monitoring, maintenance, service, and operational support. It can be designed so the customer does not have to become a hydrogen expert.

The customer gets the outcome: reliable high-purity hydrogen at the point of use.

Davy handles the complexity.

Davy’s Opportunity: Hydrogen as Reliability Infrastructure

The strongest way to position Davy in this article is not as a company chasing the hydrogen economy.

It is as a company building practical hydrogen infrastructure where hydrogen is already needed.

That distinction matters.

There is a lot of noise in hydrogen. Many proposed hydrogen applications are still waiting for economics, policy, infrastructure, or adoption. But generator cooling is different. It is an existing industrial use case tied directly to the reliability of power generation.

Davy’s on-site model fits because it addresses four real customer pain points.

First, supply reliability. Customers using hydrogen-cooled generators need hydrogen

available when they need it. On-site production reduces dependence on delivered supply.

Second, purity control. Generator performance and safety depend on maintaining proper hydrogen purity and monitoring gas quality. On-site systems can be integrated with purity monitoring and plant operating procedures.

Third, logistics reduction. Delivered hydrogen introduces trucking, handling, scheduling, storage, and emergency-delivery issues. On-site production simplifies the supply chain.

Fourth, scalability. As data centers, utilities, and self-generating customers add more power assets, modular on-site hydrogen can scale with the need.

This is the practical side of hydrogen that often gets missed.

Hydrogen is not only a fuel for the future.

Hydrogen is also an industrial utility gas that already supports the electric power system.

And as power becomes more valuable, the infrastructure around power becomes more valuable too.

The Bigger Point

The next phase of the hydrogen economy may not begin with millions of hydrogen cars or massive national pipeline networks.

It may begin with practical, local, high-value applications where hydrogen already solves a problem.

Generator cooling is one of those applications.

It is technical. It is not flashy. It is not usually part of the public conversation. But it sits directly inside one of the most important machines in the power system: the generator.

For data centers, on-site hydrogen can support mission-critical power reliability.

For utilities, it can reduce a small but meaningful operational risk.

For self-generating customers, it can simplify a necessary input to power independence.

And for Davy, it represents exactly the kind of market opportunity that makes sense: local hydrogen production, placed close to demand, solving a real industrial problem, without waiting for the entire hydrogen economy to arrive.

That may be the most important lesson.

Hydrogen does not need to do everything tomorrow to matter today.

Sometimes the best hydrogen market is not the biggest one.

It is the one where the customer already needs hydrogen, already values reliability, already understands the cost of downtime, and is ready for a better supply model.

Electric generator cooling is one of those markets.

And as data centers and utilities race to build the next generation of power infrastructure, the companies that solve these quiet reliability problems may become far more important than the headlines suggest.