Green Hydrogen Explained: How It Works, Where It Helps and Why It Is Hard to Scale

Green hydrogen is often presented as one of the fuels of a low-carbon future.

The basic idea is appealing. Use renewable electricity to split water into hydrogen and oxygen. Store the hydrogen. Move it where energy or an industrial feedstock is needed. When hydrogen is used in a fuel cell, the main direct product is water rather than carbon dioxide.

That description is chemically correct.

It is also incomplete.

Hydrogen has to be manufactured, compressed or otherwise stored, transported and eventually converted into something useful. Every step requires equipment, infrastructure and energy. Renewable electricity used to make hydrogen could often have been used directly instead. Hydrogen is difficult to transport because it contains relatively little energy per unit of volume under ordinary conditions. And although governments and companies announced enormous projects during the early 2020s, many have since been delayed or cancelled because customers are reluctant to pay the premium for low-emissions hydrogen.

The International Energy Agency's latest assessment shows both the scale of the opportunity and the gap between ambition and reality.

Global hydrogen demand surpassed 100 million tonnes in 2025, but almost all of it remained concentrated in established industrial uses such as refining and chemicals. Low-emissions hydrogen production grew by about 20%, yet still reached only around 1 million tonnes. The IEA expects the share to exceed 1% of total production for the first time in 2026. (iea.org)

So the real question is no longer:

Can renewable electricity make hydrogen?

It clearly can.

The harder question is:

Where is converting valuable clean electricity into hydrogen worth the additional cost, infrastructure and energy loss?

That distinction turns green hydrogen from a supposed universal replacement for fossil fuels into something more useful: a specialised tool for parts of the energy system that are unusually difficult to electrify directly.

Hydrogen is an energy carrier, which changes how it should be used

Hydrogen is the most abundant element in the universe, but free molecular hydrogen is uncommon on Earth. Most terrestrial hydrogen is chemically bound into compounds such as water, methane and hydrocarbons.

To obtain usable hydrogen, energy must first be supplied.

That makes hydrogen generally an energy carrier rather than a primary source of energy.

Sunlight can generate electricity.

Wind can generate electricity.

Coal and natural gas contain chemical energy extracted from geological resources.

Hydrogen usually has to be produced using one of those other energy sources.

This distinction has important practical consequences because converting energy from one form into another is never perfectly efficient.

Suppose a renewable power plant produces electricity.

If that electricity can be sent through the grid directly into an electric motor, a heat pump or a battery, relatively few major conversion stages are required.

If the electricity is instead used to make hydrogen, energy first enters an electrolyser. The resulting hydrogen may then need compression, liquefaction or conversion into ammonia. It has to be stored and transported. If the final application needs electricity again, a fuel cell or turbine converts it back.

Energy is lost at several stages.

This does not make hydrogen a bad technology.

It creates an opportunity-cost test.

If direct electricity already performs the task cheaply and efficiently, converting electricity into hydrogen first usually needs a strong reason.

That is why battery-electric cars generally make more sense than hydrogen fuel-cell cars for ordinary passenger transport, while heat pumps are usually a more efficient way to heat buildings than producing renewable hydrogen and burning it in boilers.

The IEA similarly treats hydrogen primarily as an option for sectors in which direct electrification is unavailable, technically difficult or economically unattractive, particularly parts of heavy industry and long-distance transport. (iea.org)

This provides a useful rule:

Use electrons directly where they work well. Use hydrogen where the molecule itself provides something electricity struggles to provide.

What makes hydrogen “green” depends on how the electricity is produced

Electrolysis itself is conceptually simple.

An electrolyser uses electricity to split water into hydrogen and oxygen through electrochemical reactions. The precise reactions and internal materials depend on the type of electrolyser.

Alkaline electrolysers are commercially mature and transport hydroxide ions through an alkaline electrolyte.

Proton-exchange-membrane, or PEM, electrolysers transport hydrogen ions through a specialised membrane and can respond rapidly to changing electricity supply, which can make them attractive alongside variable wind and solar generation.

Solid-oxide electrolysers operate at much higher temperatures and can use heat as part of the process, reducing the quantity of electricity required. The U.S. Department of Energy notes that solid-oxide systems commonly operate at roughly 700–800°C, compared with much lower temperatures for PEM and alkaline systems. (energy.gov)

The important climate point is that electrolysis is not automatically low-carbon.

An electrolyser powered by electricity generated primarily from coal can indirectly cause substantial greenhouse-gas emissions.

The same equipment supplied by wind, solar, hydropower, nuclear or another genuinely low-emissions source can produce hydrogen with a much smaller carbon footprint.

This is why the popular hydrogen “colour” system can be useful but misleading.

“Grey hydrogen” generally refers to hydrogen produced from natural gas without carbon capture.

“Blue hydrogen” usually refers to fossil-based hydrogen combined with carbon capture and storage.

“Green hydrogen” usually means renewable-powered electrolysis.

Other colours have been applied to coal-based production, nuclear-powered production and additional pathways.

There is no universal scientific law governing these labels.

Two projects both marketed as “green” can produce different lifecycle emissions depending on the electricity source, whether the renewable generation is additional to what the grid would otherwise have had, the hours during which electrolysers operate, and the energy required for compression, storage and transport.

Fossil-based hydrogen with carbon capture also varies depending on capture performance and upstream methane emissions.

This is why policy is increasingly moving toward actual emissions intensity rather than colour alone.

A clean-sounding label is not a substitute for measuring the greenhouse gases produced across the supply chain.

The first hydrogen challenge is cleaning up the hydrogen industry that already exists

Hydrogen is sometimes discussed as though governments are trying to create an entirely new fuel market from nothing.

In reality, the world already has an enormous hydrogen economy.

More than 100 million tonnes of hydrogen were consumed in 2025, according to the IEA. Almost all of that demand came from familiar industrial applications rather than futuristic hydrogen aircraft or household boilers. (iea.org)

Refineries use hydrogen to process petroleum.

Ammonia plants use hydrogen to manufacture fertilisers.

Methanol production requires hydrogen.

Some existing iron-production processes also consume it.

Most of today's hydrogen is still produced using unabated fossil fuels.

That suggests a relatively straightforward first priority:

replace high-emissions hydrogen where industry already needs hydrogen before inventing weak new uses simply to create demand.

The technical substitution can be simpler because the downstream industrial process already expects hydrogen.

An ammonia plant does not need to be persuaded that hydrogen is useful. Hydrogen is already part of the chemistry.

The problem is producing the molecule with much lower emissions at a price the fertiliser industry can afford.

The same logic applies to refineries, although the long-term role of refining itself changes as petroleum demand evolves.

This existing demand is also becoming where actual low-emissions hydrogen investment is concentrated. The IEA's 2026 review estimates that projects which have reached at least final investment decision could supply around 2.5 million tonnes of low-emissions hydrogen to refining and industrial facilities by 2030. (iea.org)

That is much more concrete than assuming millions of households will suddenly replace efficient electric technologies with hydrogen appliances.

Steel, shipping and aviation show where hydrogen becomes more interesting

Beyond existing industrial hydrogen demand, the strongest future applications are generally those in which direct electrification faces real physical or chemical limitations.

Primary steel production is an important example.

Traditional blast furnaces use coke derived from coal not only as an energy source but as a chemical reducing agent that helps separate oxygen from iron ore.

Hydrogen can perform part of that chemical role in direct-reduced-iron processes.

If the hydrogen is produced with low emissions and the remaining steelmaking electricity is also low-carbon, hydrogen-based direct reduction could substantially lower emissions from primary steel production.

The economics remain difficult because conventional steelmaking infrastructure is established, hydrogen is expensive, and building new direct-reduction and electric-furnace capacity requires large capital investment.

But this is exactly the kind of use where hydrogen has a reason to exist beyond merely carrying energy.

The molecule itself performs useful chemistry.

Shipping presents a different problem.

Large ocean-going vessels require enormous quantities of onboard energy. Batteries are becoming useful for ferries and some shorter routes, but carrying enough batteries for long intercontinental voyages can become difficult because of weight, volume and recharging constraints.

Hydrogen can be converted into ammonia, methanol or synthetic hydrocarbons, creating fuels with characteristics more suitable for long-distance transport.

Those extra conversions consume additional energy, and ammonia introduces toxicity and combustion challenges of its own.

But long-distance shipping is precisely the sort of sector where accepting conversion losses may make sense because the alternative is not simply plugging a ship into the electricity grid throughout an ocean crossing.

Aviation presents an even more severe energy-density constraint.

Direct hydrogen aircraft may have specialised future roles, but much current interest focuses on using low-emissions hydrogen to manufacture synthetic aviation fuels.

Again, this is expensive and energy-intensive.

The case for doing it comes from the difficulty of directly electrifying long-distance aircraft, not from hydrogen being intrinsically superior to electricity.

The same reasoning applies to some high-temperature industrial processes and potentially some forms of long-duration or seasonal energy storage.

Hydrogen's strongest applications appear where alternatives narrow.

That is a healthier deployment strategy than beginning with hydrogen and searching for as many uses as possible.

Cost is still the central commercial barrier

The chemistry of electrolysis is not the main reason green hydrogen projects struggle to reach construction.

The difficult part is economics.

Electricity is usually the dominant operating input.

An electrolyser therefore needs access to sufficiently cheap low-carbon electricity, enough operating hours to justify its capital cost and customers willing to pay for the resulting hydrogen.

These requirements can conflict.

A solar project may produce exceptionally inexpensive electricity in a sunny location during daylight hours. But if the electrolyser operates only while the sun is available, expensive equipment sits idle for much of the day.

Adding wind generation can increase utilisation.

Grid electricity can increase it further.

Storage can smooth production.

But every addition changes the economic and emissions calculation.

Electrolyser cost itself also remains important.

Installed electrolysis capacity more than doubled in 2025 to exceed 4 GW, driven substantially by major Chinese projects, and more than 2.5 GW of additional capacity was under construction for operation in 2026 when the IEA completed its latest assessment. (iea.org)

That is significant growth.

It is still tiny compared with the scale of the global hydrogen market.

And although manufacturing scale can reduce electrolyser prices, a hydrogen production facility consists of much more than the electrolyser stack. It needs power electronics, water treatment, compression, storage, controls, electrical infrastructure and often pipelines or conversion facilities.

Financing costs matter enormously because these are capital-intensive projects.

So does utilisation.

So does the price the customer is prepared to pay.

The IEA's 2026 cost analysis finds that, without policy support, the hydrogen cost that many major industrial users could economically tolerate is generally below USD 2 per kilogram. Current low-emissions production often remains above the cost at which incumbent fossil-based processes can compete. (iea.org)

This produces the industry's central commercial problem.

Producers hesitate to build because they do not have customers committed to paying the premium.

Customers hesitate to sign long contracts because they expect future hydrogen prices to fall.

Both sides wait.

This is why offtake agreements have become so important.

A project can have cheap renewable energy, an excellent electrolyser and government enthusiasm, yet remain unfinanceable if lenders cannot see who will purchase the hydrogen at a price that supports repayment.

Only around one-fifth of newly signed low-emissions hydrogen offtake volumes in 2025 were backed by firm contractual commitments. (iea.org)

The bottleneck is therefore increasingly not:

“Can somebody build the technology?”

It is:

“Who is willing to sign a bankable contract to buy the output?”

Storage, transport and water turn hydrogen into an infrastructure problem

Producing hydrogen is only the beginning.

Hydrogen contains a large amount of energy per unit of mass, but very little per unit of volume at ordinary pressure.

That makes storage and transport difficult.

One solution is compression.

Another is liquefaction at approximately –253°C.

Another is converting hydrogen into ammonia or another hydrogen-containing molecule that can be shipped using different infrastructure.

Pipelines can move hydrogen efficiently where large volumes and suitable routes justify the capital cost.

But existing gas infrastructure cannot automatically be assumed to work unchanged.

Hydrogen molecules are small and can leak through materials more readily than natural gas. Hydrogen can also contribute to embrittlement in certain metals, potentially weakening equipment depending on material, pressure and operating conditions.

Some natural-gas pipelines can be repurposed.

Others require modification or replacement.

The scale of infrastructure still committed is modest relative to announcements. The IEA reports more than 40,000 kilometres of hydrogen pipelines announced for operation by 2035, but only about 9% of that length was operational or backed by committed investment in its 2026 assessment. (iea.org)

International transport can be even more demanding.

If pure hydrogen must ultimately be delivered after shipping, the IEA estimates that liquefaction or carrier conversion and reconversion can require more than 10 kWh of energy per kilogram of hydrogen, equivalent to over 30% of the hydrogen's own energy content, while imposing minimum transport-related costs around USD 2 per kilogram in the pathways it assessed. (iea.org)

This explains why early hydrogen development is often concentrated in industrial clusters.

If an electrolyser, ammonia plant, steel plant, storage facility and port can be located relatively close together, much of the transport problem disappears.

This can be far more sensible than assuming hydrogen will immediately become a globally traded commodity transported everywhere in the same way oil is today.

Water matters locally more than globally

Electrolysis also requires water.

Water is not a minor optional input; hydrogen atoms have to come from somewhere.

The reaction itself consumes purified water, while facilities can require additional water for treatment and cooling depending on design.

At the scale of the global water system, hydrogen production is unlikely to compete with agriculture as a dominant source of water demand.

But global averages are the wrong way to assess an individual project.

Many of the world's best solar resources occur in dry regions.

A giant electrolyser placed in an already water-stressed inland location can therefore create a serious local planning problem even if hydrogen remains a small share of global water consumption.

Coastal projects can use desalination.

That adds cost, electricity use and a requirement to manage concentrated brine responsibly.

The relevant question is therefore not:

“Does hydrogen use too much water?”

It is:

“Does this project have a sustainable water supply after accounting for the needs of households, agriculture and ecosystems in this particular location?”

That is an infrastructure and governance question rather than a reason to reject electrolysis everywhere.

The hydrogen boom has entered its reality-check phase

The early 2020s produced spectacular hydrogen announcements.

Governments published strategies.

Developers proposed enormous renewable-hydrogen hubs.

Ports planned ammonia terminals.

Countries with strong wind or solar resources announced ambitions to become major hydrogen exporters.

The resulting project pipeline looked enormous.

Actual investment decisions have been much slower.

The IEA's Global Hydrogen Review 2025 had already cut potential low-emissions hydrogen production from announced projects for 2030 from 49 million tonnes per year to 37 million tonnes because of cancellations and delays.

The 2026 review reduces it again.

The announced pipeline that could theoretically be available by 2030 has fallen to about 27 million tonnes per year. Projects already committed represent only around 4.3 million tonnes, with another roughly 2 million tonnes potentially achievable if projects with stronger prospects reach investment decisions soon. (iea.org)

That does not mean the hydrogen industry has collapsed.

Nearly USD 7 billion was invested in low-emissions hydrogen projects in 2025, almost twice the previous year's level, and the IEA expects investment to approach USD 10 billion in 2026. (iea.org)

The more accurate interpretation is that the sector has moved from a period when announcements were relatively easy into a phase where projects must prove they can actually work commercially.

A press release does not need a customer.

A final investment decision does.

Projects now need:

cheap low-emissions energy;

permitting;

water;

electrolysers;

financing;

transport or nearby demand;

certification;

and customers willing to sign long-term contracts.

Many announced developments will fail those tests.

That is normal in an emerging industrial sector.

The mistake would be counting every announced megawatt as though the project were already built.

Green hydrogen matters most when it is used selectively

Hydrogen debates often divide into two exaggerated positions.

One presents green hydrogen as the molecule that will power cars, heat homes, store renewable electricity, fuel aircraft, run ships, manufacture steel and replace natural gas across the economy.

The other concludes that conversion losses and high costs make hydrogen fundamentally pointless.

Neither is particularly useful.

Green hydrogen has genuine disadvantages.

It requires large quantities of electricity.

Electrolysers cost money.

Compression and transport cost energy.

International trade can be difficult.

Public infrastructure remains immature.

Renewable hydrogen is still more expensive than conventional fossil-based hydrogen in most regions.

And many attractive-sounding applications have a simpler electric alternative.

But none of those facts eliminates the need for a low-emissions molecule in industries where molecules already perform essential chemical or energy functions.

The strongest near-term use is arguably the least glamorous:

replace the dirty hydrogen already being used.

After that come sectors where direct electrification becomes difficult: parts of primary steelmaking, fertiliser production, shipping and aviation fuels, certain chemical processes and potentially selected forms of long-duration storage.

That hierarchy matters because renewable electricity itself is valuable.

Every megawatt-hour sent into an electrolyser cannot simultaneously be used to replace fossil electricity on a grid, charge an electric vehicle or run a heat pump.

Hydrogen therefore has to justify the conversion.

Where a battery or electric motor performs the job easily, it often will not.

Where industrial chemistry requires hydrogen, or where storing and transporting energy as molecules provides a real system advantage, the calculation changes.

This is the mature role for green hydrogen in the energy transition.

Not:

“Hydrogen everywhere.”

And not:

“Hydrogen nowhere.”

Instead:

Use low-emissions hydrogen where its particular properties solve a problem that direct electricity cannot solve as well.

That is a narrower role than some of the early hydrogen-boom rhetoric promised.

It may also be a much more important one.

The world's existing hydrogen industry already consumes more than 100 million tonnes of the molecule each year and remains overwhelmingly fossil-based. (iea.org)

Cleaning up even that existing demand would require an enormous expansion of low-emissions production.

Steel, fertilisers, shipping fuels and other difficult sectors could expand the market further.

The challenge is therefore not finding every possible place hydrogen could technically be used.

There are plenty.

The challenge is directing scarce low-carbon electricity, capital and infrastructure toward the applications where using hydrogen produces enough climate and industrial value to justify the losses required to make it.

Green hydrogen is not a primary energy source waiting to replace oil, gas and electricity.

It is an energy carrier and industrial feedstock.

Once that distinction is taken seriously, its future becomes easier to understand.

Hydrogen will probably matter most not where it can replace electrons, but where electrons alone are not enough.