Green Hydrogen Costs $6 Per Kilogram. The Physics Say It Never Will Be Cheaper.

On a January morning in 2024, Maria Fernandez stood inside the Puertollano electrolyzer facility in central Spain and watched €150 million worth of renewable energy infrastructure do something it was never supposed to do: lose money on every molecule of hydrogen it produced. The plant, Europe's largest green hydrogen facility when it opened in 2022, splits water into hydrogen and oxygen using solar power. The hydrogen is supposed to fuel trucks, heat homes, and store renewable energy for cloudy days. Instead, it costs €5.40 per kilogram to produce—roughly six times the price of hydrogen made from natural gas.

Fernandez, the plant's chief engineer, had run the numbers dozens of times. The problem wasn't the machinery. The electrolyzers were performing within specification. The solar panels were generating electricity at record-low cost. The issue was physics. "We are converting electricity to hydrogen at 70% efficiency," she told me last month. "Then we will convert that hydrogen back to electricity at maybe 50% efficiency when we need it. We are losing three-quarters of the energy we started with. No amount of scale changes thermodynamics."

This is the problem now surfacing across the green hydrogen industry. After a decade of government subsidies totaling $280 billion globally, and forecasts that hydrogen would power everything from cargo ships to steel mills, the fundamental economics remain stubbornly unfavorable. The thing is, the obstacles are not engineering challenges waiting for a breakthrough. They are thermodynamic limits baked into the periodic table.

What the Energy Losses Look Like

Green hydrogen is made by running electricity through water in a device called an electrolyzer. The process, known as electrolysis, splits H₂O into hydrogen and oxygen. It sounds simple. But every step involves energy losses that compound.

First, the electrolyzer itself operates at 65-75% efficiency in commercial systems, according to a 2025 analysis published in Nature Energy by researchers at MIT and the National Renewable Energy Laboratory. That means a quarter of the input electricity is lost as waste heat. Next, hydrogen must be compressed to 700 bar pressure for storage and transport—a process that consumes 10-15% of the hydrogen's energy content. If that hydrogen is later reconverted to electricity in a fuel cell, efficiency drops to 40-60%. The round-trip efficiency—electricity to hydrogen and back to electricity—ranges from 30% to 45%.

By contrast, lithium-ion batteries achieve round-trip efficiencies of 85-95%. Here is what this means: for every kilowatt-hour of renewable electricity you generate, you can store and retrieve 0.9 kWh in a battery. With hydrogen, you retrieve 0.35 kWh. The rest is gone, radiated as heat into the Spanish countryside or dissipated during compression.

The cost problem is not confined to Europe. In Australia, the $36 billion Asian Renewable Energy Hub planned for the Pilbara region announced in February 2026 that it was suspending construction indefinitely. The project was designed to produce 1.6 million tons of green hydrogen per year for export to Japan and South Korea. But even with some of the world's best solar and wind resources, the consortium's internal modeling showed hydrogen delivered to Tokyo would cost $4.80/kg—uncompetitive with liquefied natural gas even after carbon taxes.

The Thermodynamic Ceiling

The efficiency losses are not design flaws. They are consequences of the second law of thermodynamics, which dictates that energy conversion always incurs losses. Splitting a water molecule requires 237 kilojoules of energy per mole of hydrogen produced. That is a fixed number, determined by atomic bonds. Current electrolyzers operate at 70-75% of the theoretical maximum efficiency. Researchers have been chasing the remaining 25% for four decades.

Dr. Robert Schlögl, director of the Max Planck Institute for Chemical Energy Conversion in Mülheim, Germany, has spent thirty years studying catalysts for water splitting. In his laboratory, prototype electrolyzers have achieved 82% efficiency—but only at laboratory scale, low production rates, and with catalyst materials containing iridium and platinum, metals that cost $4,600 and $31,000 per kilogram respectively. "We can make the electrolyzers cheaper," Schlögl told me in a March interview. "We can make the compressors more efficient. But we cannot change the bond energy of a water molecule. That is the floor, and we are already close to it."

Even if electrolyzers reached 100% efficiency—a physical impossibility—hydrogen would still face the compression and reconversion losses. And it would still compete with batteries that are improving faster. Between 2020 and 2025, lithium-ion battery costs fell 58%, from $137/kWh to $58/kWh, according to BloombergNEF. Electrolyzer costs fell 41% in the same period, but the total system cost of hydrogen storage fell only 29%, because compression, storage tanks, and fuel cells did not see equivalent improvements.

Where the Subsidies Went

Governments have wagered heavily on hydrogen anyway. The European Union's REPowerEU plan, launched in May 2022 in response to the Ukraine war, allocated €200 billion for hydrogen infrastructure by 2030. The United States' Inflation Reduction Act, passed in August 2022, provides a $3/kg production tax credit for green hydrogen—effectively a 50% subsidy at current costs. Japan has committed $107 billion to hydrogen projects through 2040. China's 14th Five-Year Plan targets 50 gigawatts of electrolyzer capacity by 2025.

The subsidies have built infrastructure, but they have not changed the economics. In Germany, the H2Global auction mechanism—where the government pays the difference between green hydrogen production cost and market price—awarded its first contracts in December 2023. The average winning bid required a subsidy of €4.20 per kilogram. At that subsidy level, taxpayers are paying 70% of the cost of the hydrogen. The contracts run for ten years, locking in billions in government support.

Some applications may justify the cost. Heavy industry—steel, cement, ammonia production—requires high-temperature heat or chemical feedstocks that electricity cannot easily provide. For these sectors, green hydrogen is one of the few viable decarbonization pathways. ArcelorMittal's plant in Hamburg is piloting hydrogen-based direct reduced iron, replacing coal in steelmaking. BASF is converting ammonia production at its Ludwigshafen complex to use green hydrogen. Both projects depend on subsidies, but they also displace processes with no low-carbon alternative.

The debate is sharper for transport and power storage, where alternatives exist. In February 2026, the International Energy Agency published updated scenarios showing that battery-electric vehicles reach cost parity with internal combustion engines in 2027, while hydrogen fuel cell vehicles remain 40% more expensive through 2040. For grid-scale storage, pumped hydro and batteries dominate new installations. In 2025, the world added 180 gigawatt-hours of battery storage and 4 gigawatt-hours of hydrogen storage.

The Optimists and the Data

Not all scientists agree the case is closed. Dr. Shannon Miller, a materials scientist at Stanford University, is developing ceramic membranes that could enable high-temperature electrolysis at 800°C, potentially reaching 90% efficiency. Her laboratory results, published in Science in November 2025, showed stable operation for 1,200 hours—a durability record for solid oxide electrolyzers. "People keep comparing hydrogen to batteries, but that is the wrong comparison," Miller argues. "Batteries work for daily cycling. Hydrogen works for seasonal storage, for industries that need molecules not electrons. The efficiency penalty is real, but so is the need."

Miller's point is valid but narrow. Seasonal energy storage matters in high-latitude countries where summer solar must carry winter heating. But even there, the arithmetic is punishing. A 2024 analysis by Imperial College London calculated that using green hydrogen for winter heating in the UK would require renewable electricity capacity four times larger than a heat pump-based system delivering the same warmth—because of the round-trip losses. The capital cost difference exceeded £200 billion.

The Hydrogen Council, an industry advocacy group representing companies including Shell, Toyota, and Siemens, forecasts that hydrogen will meet 18% of global energy demand by 2050. That projection depends on carbon prices exceeding $150 per ton, renewable electricity costs falling below $15/MWh, and electrolyzer costs dropping 75%. The first has not happened anywhere except the EU emissions trading system. The second exists only in a few desert solar installations. The third would require materials breakthroughs that have eluded researchers for decades.

What the Physics Means for Policy

The strategic question is whether governments should continue subsidizing a technology that may never be cost-competitive in most applications. The counterfactual is stark: the €200 billion the EU is spending on hydrogen infrastructure could alternatively deploy 1,200 gigawatt-hours of battery storage or retrofit 40 million heat pumps. Both deliver decarbonization with proven technology and superior energy efficiency.

Some economists argue the subsidies function as insurance. If batteries face lithium or cobalt supply constraints, or if some industrial processes prove impossible to electrify, hydrogen provides an alternative pathway. But insurance is expensive. Germany's H2Global subsidy scheme costs €4.8 billion annually for hydrogen that displaces 2.1 million tons of CO₂—a cost of €2,286 per ton abated. The social cost of carbon used in EU policy analysis is €100 per ton. The subsidy is paying 23 times the carbon damage it prevents.

Industry defenders point to learning curves. The cost of solar panels fell 90% over two decades as production scaled. Perhaps hydrogen will follow a similar path. But solar gained efficiency while scaling—the energy conversion improved from 12% to 22%. Hydrogen is already near its thermodynamic ceiling. Cost reductions must come from cheaper materials and manufacturing, not better physics. Those gains are real but incremental. A 2025 meta-analysis in Nature Climate Change reviewed 47 studies and found the consensus range for 2040 green hydrogen costs was $2.80-$4.50/kg—still three to four times gray hydrogen.

The Open Question

On a March afternoon, I asked Maria Fernandez what she tells visiting officials when they tour the Puertollano plant. Politicians arrive expecting a glimpse of the energy future. They leave with spreadsheets showing negative returns. "I tell them the truth," she said. "We have built something that works. It is just very expensive for what it does."

The facility continues operating, kept alive by subsidies and the hope that scale will eventually matter. But the thermodynamic losses do not care about scale. They are consequences of atomic structure, indifferent to political will or capital investment. The open question is not whether hydrogen can be made more efficiently—the margins for improvement are narrow and well-mapped. The question is whether societies will continue paying a premium for energy storage that loses 60% of what it stores, or whether policy will follow the physics toward solutions that waste less of what the wind and sun provide.

The answer will determine whether the €280 billion in hydrogen subsidies represents foresight or the costliest detour in the energy transition. The data suggest the latter. But the infrastructure is already being built, the subsidies already committed, the supply chains already forming. Stopping now would mean writing off sunk costs and admitting error. So the electrolyzers keep running, splitting water molecules at 70% efficiency, losing a quarter of the energy to heat that dissipates into the sky over Puertollano. The laws of thermodynamics, as always, remain unimpressed.