“Earth’s crust has produced enough hydrogen over the past billion years to meet our current energy needs for 170,000 years,” according to Chris Ballentine, professor and chair of geochemistry at the University of Oxford. That mind-boggling calculation, published in Nature Reviews Earth and Environment, is rewriting the search for potentially one of the Earth’s most impactful as-yet-unexploited energy sources: naturally occurring subsurface hydrogen.
For many decades, the general assumption was that hydrogen’s tiny molecular size and high reactivity made it impossible for it to build up in commercially significant quantities. Millions of oil and gas wells drilled around the globe only sporadically picked up more than trace levels. But discoveries in Mali, Albania, and the U.S. Midwest in recent times have destroyed that concept, revealing that if the right geological conditions exist, hydrogen can be formed, trapped, and stored over millions of years.
The new study condenses those conditions into an “ingredient list” to study. Three ingredients are necessary: a source of hydrogen, reservoir rocks to hold it, and impermeable caps to confine it. Of a dozen known natural generation mechanisms, the most productive is serpentinization a geochemical reaction that involves water seeping into iron-rich ultramafic rocks like olivine, oxidizing the iron and liberating hydrogen. Laboratory experiments demonstrate that serpentinization at 200–320°C can produce 182 billion cubic feet of hydrogen per cubic kilometer of rock per year, orders of magnitude more rapidly than the rate of hydrocarbon generation. Radiolysis, decomposition of water by natural radioactivity in uranium-, thorium-, or potassium-containing minerals, also generates hydrogen on geologic timescales, but at much more modest rates.
Once produced, hydrogen has to migrate into permeable, porous reservoirs sandstones, fractured carbonates, or weathered basalts and be capped by impermeable seals like clays, evaporites, or unfractured volcanic flows. The Midcontinent Rift in Kansas is a good example of this geometry. Created 1.1 billion years ago, it hosts thick basalt units that can produce hydrogen from water–rock interaction, capped by sedimentary sequences that may serve as reservoirs and seals. Extraction firms such as Koloma and HyTerra are drilling up this ancient scar, focusing on structures close to the Nemaha Ridge where ancient oil and gas wells have measured hydrogen concentrations up to 96 percent.
Other encouraging environments are ophiolite complexes pieces of oceanic crust pushed onto continents in which ultramafic rocks come into contact with meteoric water. In 2024, a giant hydrogen reservoir was discovered in Albania’s Bulqizë chromite mine, venting some 200 tons of hydrogen per year from a faulted Jurassic ophiolite massif. Large igneous provinces and Archaean greenstone belts, over 3 billion years old in some cases, also offer good combinations of source rocks, heat flow, and tectonic structures.
Not all settings are compatible. As co-author Barbara Sherwood Lollar of the University of Toronto observed, “We know for example that underground microbes readily feast on hydrogen,” so biologically active areas may consume accumulations before they are found. High-silica settings can also inhibit serpentinization, cutting returns.
Measuring subsurface hydrogen involves modifying petroleum and geothermal exploration equipment. Surface seep surveys, soil gas sampling, and geophysical imaging can detect migration routes and possible traps. Downhole mud gas logging, utilized in Kansas drilling operations, gives on-site hydrogen readings, while laboratory isotopic analysis can verify origin and maturity. The U.S. Geological Survey has constructed a national prospectivity map, indicating areas from the Great Plains to the Eastern Seaboard with favorable set-ups of hydrogen sources, reservoirs, and seals.
The stakes are higher than geology. More than 95 percent of industrial hydrogen today is made from natural gas through steam methane reforming, releasing 10–12 kilograms of CO₂ per kilogram of hydrogen. Even so-called “blue” hydrogen with carbon capture has 3–5 kg CO₂e/kg H₂. In comparison, geologic hydrogen produced from direct production in reservoirs may be on a carbon footprint as low as 1 kg CO₂e/kg H₂, assuming production and handling reduce methane leakage and other upstream emissions. Since hydrogen demand is expected to double to 200 million tonnes a year by 2030, low-carbon sources will be essential in the decarbonization of applications such as steelmaking, fertilizer manufacturing, and heavy transport.
The two-part challenge now is to demonstrate that accumulations are big and accessible enough for commercialization, and to develop more advanced engineering to recover and process the gas safely. The embrittling effect of hydrogen on steel requires sensitive choice of materials for wells and pipelines. Separation from associated gases, underground storage, and integration with existing hydrogen infrastructure will need more innovation.
Nevertheless, the intersection of geological understanding, detection technology, and policy momentum is speeding up the field. As Ballentine has said, Earth’s crust produces plenty of hydrogen it is now a question of following the ingredient list to find it. For energy strategists, geologists, and investors, that list might be the guide to one of the century’s biggest low-carbon energy breakthroughs.
