Clean hydrogen is the missing piece in decarbonizing steel, shipping, and chemicals. Fusion’s intense heat could produce it at scale without any carbon emissions.
Hydrogen is often called the fuel of the future, but today roughly 95 percent of the world’s hydrogen is made from natural gas through a process called steam methane reforming — a method that emits about 10 tonnes of carbon dioxide for every tonne of hydrogen produced.1 Decarbonizing hydrogen production is essential for cleaning up industries like steelmaking, ammonia synthesis, and long-haul shipping, where direct electrification is impractical. Fusion energy could offer a uniquely powerful path to carbon-free hydrogen at industrial scale.
The hydrogen industry uses a color-coding system to describe production methods. Grey hydrogen comes from natural gas with no carbon capture. Blue hydrogen adds carbon capture to the same process. Green hydrogen uses renewable electricity to split water via electrolysis. Each has limitations: grey is dirty, blue still leaks methane and CO2, and green competes with the grid for renewable power that is already in short supply.
Fusion-produced hydrogen would be truly zero-carbon at the point of production, with no intermittency problems and no upstream methane emissions. Some analysts have begun referring to it informally as “gold hydrogen” — produced by a zero-carbon, high-capacity-factor thermal source.
There are two primary pathways for using fusion energy to produce hydrogen.
High-temperature electrolysis (HTE): Conventional electrolysis splits water into hydrogen and oxygen using electricity. At higher temperatures (700–1,000°C), the thermodynamic energy required drops significantly, meaning less electricity is needed per kilogram of hydrogen. A fusion plant could supply both the electricity and the high-grade heat, pushing overall conversion efficiencies above 50 percent — well beyond what low-temperature electrolysis achieves.2
Thermochemical water splitting: These are chemical cycles that use heat alone (no electricity) to decompose water. The sulfur-iodine (S-I) cycle and the copper-chlorine (Cu-Cl) cycle are the most studied. The S-I cycle operates at temperatures around 850–950°C, which is within reach of fusion blanket coolant systems, particularly those using molten salts or high-pressure helium.3
Global hydrogen demand is currently about 90 million tonnes per year and is projected to grow to 150–600 million tonnes by 2050 as new applications come online. Producing that volume through electrolysis alone would require an enormous buildout of renewable generation capacity — several thousand gigawatts dedicated solely to hydrogen, on top of what the grid already needs.4
A single 1-GW fusion plant, dedicating a portion of its thermal output to high-temperature electrolysis, could produce roughly 100,000–150,000 tonnes of hydrogen per year. A fleet of such plants could make a meaningful dent in global demand without competing with the electricity grid for renewable capacity.
The most compelling deployment scenario places fusion-hydrogen plants at industrial clusters where hydrogen demand is concentrated: oil refineries transitioning to clean feedstock, steel mills replacing coal with hydrogen-based direct reduction of iron ore, and ammonia plants supplying the fertilizer industry.
Co-locating fusion plants with these consumers eliminates the need to transport hydrogen over long distances — a major cost and safety challenge, since hydrogen is difficult to store and pipe compared to natural gas.
Fusion-driven hydrogen production inherits all the challenges of fusion energy itself: demonstrating sustained net energy gain, qualifying materials for neutron environments, and achieving competitive capital costs. Beyond that, the hydrogen production systems — particularly thermochemical cycles — require materials that can withstand corrosive chemical environments at very high temperatures for years.5
The timeline is necessarily linked to fusion commercialization. Most roadmaps place the first commercial fusion plants in the 2035–2045 window. Hydrogen co-production could be integrated from the outset or retrofitted to early plants as the technology matures.