Let’s Talk About Hydrogen: Is It the Future of Power Generation or Just Expensive Backup?

Hydrogen can make electricity. That part is not controversial anymore.

The harder question is why a utility would want to use it.

If the goal is to move solar power from late afternoon into the evening, batteries are usually the cleaner and more efficient answer. NREL’s current utility-scale battery assumptions use roughly 85% round-trip efficiency, and the technology is now broadly modeled in durations from 2 to 10 hours. EIA data has also shown that utility batteries built through 2020 averaged about 3 hours of duration, while four-hour systems became a key benchmark for reliability value in organized markets.

If the goal is dispatchable thermal power at the lowest near-term cost, natural gas still has the incumbent advantage. If the goal is firm carbon-free generation, utilities may look first to nuclear, geothermal, or expanded transmission. And if the goal is simply to avoid building more generation, demand response and load flexibility can often solve part of the problem before a new power plant is ever needed.

So why does hydrogen keep showing up in utility planning decks, pilot programs, turbine roadmaps, and decarbonization scenarios?

Because the electric grid is not judged on ordinary days. It is judged on the hardest ones.

Utilities are not looking for a miracle fuel. They are trying to solve a brutal equation: keep the lights on, cut emissions, contain cost, reduce long-term policy risk, and avoid building assets that look prudent today but stranded tomorrow. That is why hydrogen remains relevant. Not because it is obviously superior, but because it may solve a narrow class of problems that become more important as the grid gets cleaner, tighter, and more weather-exposed.

Hydrogen enters the power conversation in several ways. It can be blended into natural gas and burned in turbines. It can be the primary fuel in future hydrogen-capable combustion systems. It can be converted directly into electricity through fuel cells. Or it can function as stored energy: electricity makes hydrogen through electrolysis, the hydrogen is stored, and later it is turned back into electricity when the grid needs it. DOE explicitly frames hydrogen as a possible enabler of long-duration energy storage and grid flexibility, especially in systems with rising shares of renewable and nuclear generation.

That sounds powerful. And strategically, it is. But physically and economically, there is a catch.

Hydrogen is usually a bad answer to the wrong problem.

The reason is efficiency. Batteries may lose some power charging and discharging, but hydrogen loses energy in more places: first when electricity is converted to hydrogen, then in compression, storage, and handling, and then again when the hydrogen is converted back into electricity. At NREL’s ARIES Flatirons work, a PEM electrolyzer plus PEM fuel cell system showed a full-system round-trip efficiency of about 28.3% at full power, with a higher case around 35.1% under lower-power solar-coupled conditions. That is a dramatic gap compared with battery storage.

This single fact explains most of the hydrogen-for-power debate.

If a utility is solving a short-duration balancing problem, hydrogen is usually too lossy. If it is solving a daily cycling problem, hydrogen is usually too infrastructure-heavy. If it is solving a straightforward capacity need, hydrogen often loses on cost and complexity. That is why hydrogen should not be treated as the default future of electricity generation. In most cases, it is not the best answer.

But that does not mean it is irrelevant. It means its role is more specific and, in some ways, more interesting.

The strongest case for hydrogen begins where batteries start to struggle.

DOE’s Long Duration Energy Storage framing draws an important line: short-duration storage generally sits below 10 hours, inter-day storage spans roughly 10 to 36 hours, multi-day and week-long storage reaches 36 to 160 hours, and seasonal shifting goes beyond 160 hours. DOE has also estimated the United States may need between 225 and 460 gigawatts of long-duration energy storage by 2050 in net-zero pathways. That is a staggering system challenge, and it is one of the clearest reasons hydrogen keeps surfacing in utility strategy.

That does not mean hydrogen wins the long-duration storage market. It does mean the battleground changes once the storage problem stops being four hours and starts becoming two days, five days, or an entire seasonal mismatch between renewable production and demand.

This is where hydrogen’s weakness starts to become its advantage.

Hydrogen is inefficient, yes. But unlike batteries, it is not trying to win on round-trip efficiency alone. It is trying to compete in a different category: very large energy capacity over long periods of time. Hydrogen’s value is not that it stores electricity elegantly. It is that it can store energy for long periods without the same self-discharge profile or cost structure associated with simply scaling lithium-ion systems further and further out in duration. NREL has specifically identified hydrogen seasonal energy storage as a case that could become cost-competitive in future high-renewable systems.

There is also a second reason hydrogen matters to utilities, and it has less to do with chemistry than with legacy infrastructure.

Utilities already own thermal assets. They understand turbines. They understand dispatchable fuel. They understand spinning machinery, outage schedules, heat rates, ramp profiles, and capacity value. Hydrogen is attractive in part because it can fit, however imperfectly, into a world utilities already know how to operate.

That is why turbine manufacturers have leaned so heavily into hydrogen-capable roadmaps. GE Vernova says its installed fleet includes gas turbines with operating experience on fuels containing anywhere from 5% to 100% hydrogen by volume, and its advanced HA turbines are marketed with 50% hydrogen capability today and a pathway toward 100% hydrogen. Mitsubishi Power Americas says its JAC systems are 30% hydrogen capable, and its H-100 line is also listed as hydrogen capable up to 30%, with simple-cycle output of roughly 105 to 116 MW and combined-cycle output up to 350 MW. These are not proof of economic superiority. But they are proof that hydrogen is no longer just a lab concept in the combustion world.

That matters because utilities are not planning on a blank sheet of paper. They are planning around existing interconnections, brownfield sites, gas fleets, workforce capabilities, and political realities. A fuel that can potentially repurpose part of the thermal power stack will always get more attention than a technology that requires an entirely new operating model.

Even so, the economics are still brutal.

DOE’s Hydrogen Shot remains the sector’s most famous benchmark: reduce clean hydrogen production cost by 80% to $1 per kilogram by 2031. The very existence of that target tells you how far hydrogen still has to go. When the U.S. government’s flagship objective is essentially to make the molecule dramatically cheaper, it is an acknowledgment that cost, not theoretical usefulness, remains the central obstacle. DOE’s updated clean hydrogen commercialization work continues to emphasize that hydrogen will matter most in sectors where there are few other cost-effective decarbonization options.

That point is especially important in the power sector.

Hydrogen for electricity does not just depend on the cost of producing hydrogen. It depends on the cost of producing it, moving it, storing it, conditioning it, and then converting it back into usable power. The molecule may be the headline, but the infrastructure is the bill. DOE’s Hydrogen Program Plan explicitly describes hydrogen as an opportunity for storage, flexibility, and cross-sector integration, but that is also another way of saying hydrogen becomes more compelling when it can serve multiple purposes beyond just making electrons.

That is why one of hydrogen’s most realistic openings may be hybrid rather than standalone.

If a region already has industrial hydrogen demand, hydrogen logistics, geological storage, curtailed clean power, and a utility trying to decarbonize dispatchable capacity, the pieces start to align. The power sector alone may not justify the buildout. But the power sector as one customer in a broader hydrogen ecosystem might. This is one reason analysts and DOE alike keep treating hydrogen not as a universal answer, but as a sector-coupling tool with greatest value where multiple markets overlap.

There is also a systems-level reason hydrogen remains hard to dismiss: transmission.

In theory, many grid problems are better solved with wires than with backup fuel. A stronger transmission network can move low-cost renewable electricity over long distances, reduce congestion, diversify weather risk, and lower the need for local peaking assets. In practice, transmission in the United States is notoriously slow to build. DOE says developing and building transmission projects can take upwards of 10 years, and notes that additional transmission capacity in the last decade was added at roughly half the average rate of the prior three decades.

That does not make hydrogen better than transmission. It makes hydrogen more relevant in a world where the theoretically best solution may not arrive on time.

The same logic applies to firm low-carbon generation more broadly. Nuclear may offer a cleaner long-run answer for some regions, but timelines and capital intensity remain daunting. Geothermal may be superb where it is viable, but it is geographically constrained. Batteries are excellent where the duration requirement is manageable, but less straightforward once the grid needs multiple days of insurance. Hydrogen lives in the spaces created by those constraints.

Which brings us back to the title question: future of power generation, or just expensive backup?

The honest answer is probably closer to the second one, but with an important twist.

Hydrogen is unlikely to become the mainstream fuel of electricity generation. It is too inefficient, too infrastructure-heavy, and too expensive to displace the grid’s more obvious tools in everyday operation. It is not likely to outcompete batteries for short-duration balancing. It is not likely to beat direct electrification where direct electrification works. And it is not likely to become the first thing a prudent utility reaches for when cheaper options remain available.

But “expensive backup” undersells what the grid may eventually need.

In a deeply decarbonized system, backup is not a side issue. Backup becomes strategy.

The harder the grid leans on weather-dependent generation, the more valuable it becomes to have something that can sit quietly for long periods and then deliver power when the easy options fail. That is not a glamorous use case. It is not the kind that produces viral hydrogen headlines. But it may be the one that matters.

Utilities do not need hydrogen to win everywhere. They need it to work where other tools start to break down: when storage duration stretches from hours into days, when transmission cannot be built fast enough, when legacy thermal assets need a lower-carbon future, and when resilience starts to matter more than elegance.

That does not make hydrogen the future of power generation in the broad, triumphant sense often implied by industry hype.

It does make hydrogen a serious contender for one of the grid’s toughest jobs.

And that may be enough.

Batteries vs. Hydrogen: The Real Tradeoff

Batteries and hydrogen are often discussed as if one must replace the other. In reality, they solve different grid problems. Batteries are usually superior when the system needs fast response, high round-trip efficiency, and routine shifting of power over a few hours. Hydrogen becomes more interesting when the utility is trying to move large amounts of energy across much longer time horizons, or when the fuel also has value beyond the power sector.

That distinction matters because batteries carry real constraints of their own. Today’s utility-scale battery market is dominated by lithium-ion systems, and NREL’s 2024 ATB models those systems primarily at 2 to 10 hours of duration, with fixed O&M assumptions that include augmentation to maintain rated capacity over a 15-year lifetime. That makes batteries highly effective for daily balancing, ramping support, and peak shaving, but less natural as a solution for multi-day or seasonal storage. They are also fundamentally stationary assets once installed. Hydrogen, by contrast, can be produced centrally or locally, stored in bulk, and then used in power generation, industrial applications, or transportation depending on where the demand exists.

The environmental comparison is also more nuanced than either side of the debate usually admits. Battery supply chains depend on mineral extraction and processing, and NREL notes that life-cycle assessments have identified battery production as a significant source of environmental impacts, including upstream mining and refining burdens. At the same time, lithium-ion recycling is improving quickly. A 2025 Nature Communications study found that converting mixed-stream end-of-life batteries into battery-grade materials can reduce environmental impacts by at least 58% compared with conventional mined-material supply chains. So it is fair to say batteries are material-intensive and still face end-of-life challenges, but it is no longer accurate to describe them as a one-way environmental liability.

Hydrogen has almost the opposite profile. It is harder to use efficiently, but easier to imagine at very large storage volumes and across more use cases. Electricity cost is often the single biggest lever in green hydrogen economics; a 2025 review of electrolyzer economics found that electricity accounts for more than 60% of levelized hydrogen cost across major electrolyzer technologies. That is why low-cost off-peak power, curtailed renewable energy, or steady nuclear output can materially improve the case. DOE has emphasized nuclear-powered hydrogen as one particularly attractive pathway, noting that a single 1,000-MW nuclear plant could produce up to 150,000 tons of hydrogen per year. In that setup, off-peak nuclear generation does not have to be spilled or sold into weak power prices; it can be converted into a storable product that may serve both the power system and nearby industrial or mobility demand.

At a glance

Topic	Utility batteries	Hydrogen
Best use case	Fast response and short-duration shifting (roughly hours)	Longer-duration, bulk, or cross-sector energy storage
Efficiency	High; NREL ATB uses about 85% round-trip efficiency	Lower; NREL ARIES found about 28%-35% for electrolyzer-plus-fuel-cell round trip
Typical modeled life	About 15 years with augmentation in NREL utility-scale assumptions	Storage medium can sit for long periods; conversion equipment still has capex and maintenance needs
Mobility / delivery	Largely fixed once installed	Can be produced on-site or distributed by pipeline/truck depending on infrastructure
Environmental profile	Material- and mining-intensive, but recycling is improving rapidly	Lower direct production emissions if made from clean power, but lower end-to-end efficiency
Where it tends to win	Daily balancing, ramping, peak shaving	Multi-day or seasonal storage, fuel flexibility, repowering thermal assets

How Far Are We From the $1/kg Goal?

Quite far, especially for clean hydrogen made by electrolysis. DOE’s January 2025 Energy Earthshots report says hydrogen produced with electrolyzers in grid-connected scenarios is typically above $6 per kilogram today, while direct coupling with clean electricity can get some current cases near $5 per kilogram. NREL’s 2024 PEM electrolyzer manufacturing update similarly notes that levelized hydrogen production from steam methane reforming is still around $2 per kilogram, versus roughly $5 to $6 per kilogram from PEM electrolysis under current assumptions. That gap is exactly why the DOE target remains so aggressive and so important.

On the retail side, the spread is even wider because delivery, compression, storage, dispensing, and station utilization all matter. NREL’s transportation ATB notes that current market hydrogen prices are highly variable and that retail hydrogen prices include both production and infrastructure costs. In California, which remains the primary U.S. retail market, hydrogen pump prices climbed above $32 per kilogram in 2025 according to market tracking cited by Stillwater Associates, after already reaching record levels above $34 per kilogram in late 2024 according to S&P Global Commodity Insights. In other words, the industry is not just trying to move from $5 or $6 hydrogen to $1 hydrogen at the production gate; in transportation markets it is also trying to compress an end-user price that can be several multiples higher once the full retail chain is included.  California has several unique problems which contribute to a high cost for hydrogen and other goods.  Diesel and gasoline are approaching $8.00/gallon due to refinery closers, no pipelines from the Gulf Coast and very little use of California’s massive oil deposits. This makes all transportation very expensive in California. Additionally, the costs of electricity is 2 X the national average at $0.35/kWh. This is a result of closing power plants to the point that the state must import 25 – 30% of the electricity used in the state.  California is a worst case for the cost of Hydrogen.

That is also why low-cost power matters so much. If electricity drives more than 60% of electrolytic hydrogen cost, then cheap off-peak electricity and high utilization can move the economics dramatically. Nuclear is especially interesting here because it can provide very high capacity-factor electricity when market power prices are soft, and the hydrogen can then be consumed locally, sold regionally, or used later for power generation. The closer production, storage, and end use sit to one another, the more hydrogen starts to look like a strategic local fuel rather than an expensive transported commodity.

Selected references

·        U.S. Department of Energy, Hydrogen Shot: An Introduction and DOE hydrogen program materials on the $1/kg by 2031 target.

·        U.S. Department of Energy, Hydrogen Program Plan 2024, on hydrogen’s role in long-duration storage, flexibility, and power system integration.

·        NREL, Utility-Scale Battery Storage assumptions and Cost Projections for Utility-Scale Battery Storage, including representative round-trip efficiency and modeled system durations.

·        U.S. Energy Information Administration, on average utility battery duration and storage use cases.

·        NREL ARIES / Flatirons hydrogen system work, including full-system round-trip efficiency figures for electrolyzer-plus-fuel-cell storage.

·        DOE Long Duration Energy Storage materials, including the 225–460 GW estimate and the duration categories that distinguish short-duration, multi-day, and seasonal storage.

·        DOE transmission permitting materials and the National Transmission Needs Study, on grid expansion delays and development timelines.

·        GE Vernova materials on current hydrogen-capable gas turbine platforms and blending capability.

·        Mitsubishi Power Americas materials on hydrogen-capable turbine platforms.

·        NREL, Utility-Scale Battery Storage (2024 ATB), on duration range, 85% round-trip efficiency, and 15-year lifetime assumptions. https://atb.nrel.gov/electricity/2024/utility-scale_battery_storage

·        NREL, Electric Vehicle Lithium-Ion Battery Life Cycle Management, on recycling, reuse, and battery circularity challenges. https://docs.nrel.gov/docs/fy23osti/84520.pdf

·        NREL, A Framework for Integrating Supply Chain, Environmental, and Social Justice Factors During Early Stationary Battery Research, on upstream mining and social impacts. https://docs.nrel.gov/docs/fy24osti/86123.pdf

·        Nature Communications (2025), on industrial-scale lithium-ion battery recycling reducing environmental impacts by at least 58% versus conventional mined-material pathways. https://www.nature.com/articles/s41467-025-56063-x

·        NREL ARIES / Flatirons report, on full-system hydrogen storage round-trip efficiency of roughly 28%-35%. https://www.nrel.gov/docs/fy24osti/89231.pdf

·        DOE Energy Earthshots Initiative Report (January 2025), on current electrolytic hydrogen costs typically above $6/kg in grid-connected cases and near $5/kg in some clean-power-coupled cases. https://www.energy.gov/sites/default/files/2025-01/doe-energyearthshots-initiativereport.pdf

·        NREL, Updated Manufactured Cost Analysis for Proton Exchange Membrane Water Electrolyzers, on current hydrogen production cost benchmarks of about $2/kg for SMR and $5-$6/kg for PEM electrolysis. https://docs.nrel.gov/docs/fy24osti/87625.pdf

·        DOE, 3 Nuclear Power Plants Gearing Up for Clean Hydrogen Production, on the use of nuclear generation for hydrogen and DOE’s estimate that a 1,000-MW nuclear plant could produce up to 150,000 tons of hydrogen annually. https://www.energy.gov/ne/articles/3-nuclear-power-plants-gearing-clean-hydrogen-production

·        DOE Alternative Fuel Price Report (October 2025), noting that hydrogen data are collected but that retail hydrogen remains a limited and highly variable market. https://afdc.energy.gov/files/u/publication/alternative_fuel_price_report_october_2025.pdf

·        Stillwater Associates (May 2025), on California retail hydrogen averaging above $32/kg in 2025 year-to-date. https://stillwaterassociates.com/2025-cost-showdown-for-drivers-is-hydrogen-fuel-cheaper-than-gasoline/

·        S&P Global Commodity Insights (October 2024), on California retail hydrogen around $34.55/kg during supply disruptions. https://www.spglobal.com/energy/en/news-research/latest-news/energy-transition/100124-california-hydrogen-pump-prices-for-light-duty-vehicles-reach-new-highs