Back in the 1990s, Barbara Sherwood Lollar took a trip down the Kidd Creek mine in northern Ontario - a descent of more than three kilometers into the ancient root of North America. Her team of geochemists found water that had been trapped underground for over a billion years, and it turned out to be a cozy habitat for microbes that snack on hydrogen produced by water-rock reactions.

Decades later, Sherwood Lollar, now at the University of Toronto, dusted off that hydrogen data to see if the mine could serve as a source of zero-carbon fuel. "If we can set some smart minds into figuring out how to hook it up and use it, then we've got a win for this nascent economy," she says. Because, as everyone knows, the only thing better than discovering ancient microbial life is figuring out how to turn it into fuel.

Hydrogen fuel, while promising, typically requires more energy to produce than it contains and generates plenty of greenhouse-gas emissions in the process. But "geologic hydrogen" - the ready-made stuff underground - could change the equation. So a flurry of exploration has launched worldwide, with startups like Australia's HyTerra and Bill Gates - backed Koloma poking around the US Midwest in search of ancient oceanic rocks that might yield the gas.

The US Geological Survey estimates trillions of tons of H2 are produced within Earth's crust; recover even a fraction, and you could meet global hydrogen demand for centuries. But so far, no one has found a commercially viable reservoir. Public data is scarce as companies jockey for position, but at Kidd Creek, Sherwood Lollar and colleague Oliver Warr took a different approach: they looked at their own long-term data from 35 boreholes, each consistently releasing an average of eight kilograms of hydrogen per year. Extrapolating to the mine's 14,000 boreholes, that's about 140 metric tons of gas flowing unused out of the vents annually.

Published in PNAS, this isn't a world-changing amount, but it could power a substantial portion of the mine's operations - a modest local demonstration that geologic hydrogen can actually be used. Laurent Truche, a geochemist at the University of Grenoble Alpes in France, says the Kidd Creek results add to "the growing evidence that natural hydrogen generation and migration are genuine geological processes." His team reported in 2024 that at least 200 metric tons of hydrogen flow out of the Bulqizë chromium mine in Albania every year. The challenge, he notes, "is not proving that natural hydrogen exists, but proving that it can be produced economically and reliably at commercial scale."

But maybe we don't need the mother lode. Researchers are also exploring stimulation - injecting water, heat, or catalysts into reactive rocks to speed up hydrogen production. More than a dozen such projects are funded by ARPA-E, which aims to accelerate the reaction by a factor of 10,000, the rate at which stimulated production would be commercially viable.

A hint that this could work came from Oman, where a team drilled a one-kilometer borehole and injected 50,000 cubic meters of water. When they reopened the well months later, gas was spewing out - 90% hydrogen. "It's bubbling with gas," Jo Shannon, a geoscientist at the University of Southampton, told attendees at the European Geosciences Union conference in May. But she was careful to note that a crucial question remains: Is that hydrogen newly made through stimulation, or had it been lurking there all along? Because, as always, the universe enjoys a good practical joke.