Where did Earth's water come from? It is one of the oldest questions in planetary science, and for decades the most popular answer has been delivery from space: icy asteroids and comets crashing into the young Earth, bringing frozen water with them. Now a new study suggests a very different possibility. Earth may have manufactured much of its own water while it was still growing, in a hellish atmosphere of hydrogen gas, through simple chemistry between rock and air.
The research, published in Nature Geoscience and reported on 27 October 2026, comes from the University of New Mexico. Graduate student Susmita Garai, together with professor Zachary Sharp and Peter Olson, ran high-temperature experiments that recreated the conditions of the early Earth. Their results show that water can form in minutes when iron-bearing rock meets hydrogen gas at extreme heat. If the same process worked on the young planet, the team calculates, it could have produced hundreds of oceans' worth of water — far more than Earth holds today.
This is not yet a settled answer. It is a laboratory result, tested on small rock samples, then extrapolated to a planet forming over millions of years. But it changes how scientists think about where habitable planets get their water.
The old mystery: a wet planet in a dry neighbourhood
Earth is the only known planet covered mostly by liquid water, and that water is the reason life exists here. But the inner solar system, where Earth formed, was hot. Close to the young Sun, ice could not easily survive, and the rocky material that built Earth should have been mostly dry.
That puzzle led scientists to the delivery idea. Further out in the solar system, beyond the so-called frost line, water ice was stable. Icy pebbles, asteroids and comets from those cold regions could have carried water inward and delivered it to Earth in a rain of impacts. It is a good theory, and it probably explains at least part of Earth's water. But it has always had awkward questions attached: how much water could really survive the violence of those impacts, and does the chemistry of Earth's water match the chemistry of the proposed delivery vehicles? A striking example of how space surveys keep finding unexpected icy visitors is Euclid's first comet discovery, which showed a new telescope catching objects other surveys had missed.
The new study offers a second route that does not depend on long-distance delivery at all. Instead of water arriving from elsewhere, the ingredients for water were already present where Earth was forming: rocky material containing iron oxide, and hydrogen gas in the young planet's atmosphere. All that was needed was heat — and the early Earth had heat in abundance.
Pebble accretion: building a planet from falling stones
To understand the new idea, it helps to picture how Earth grew. One leading model of planet formation is called pebble accretion. In this picture, a planet does not grow only by giant collisions between large bodies. It also sweeps up huge numbers of small rocky particles — pebbles, centimetres to metres across — that drift through the disk of gas and dust around the young Sun.
As the growing planet becomes massive enough, it holds on to a thick atmosphere. Hydrogen was the most abundant gas in the early solar system, so that atmosphere would have been hydrogen-rich, very hot, and high-pressure near the surface. Pebbles falling through this atmosphere would plunge into extreme conditions: temperatures over a thousand degrees, and a constant bath of hydrogen gas.
This is the setting the New Mexico team recreated in the laboratory. And it is worth noting how central hydrogen is to the early universe in general — hydrogen is not only the fuel of stars but the most common element in existence, as shown by recent work such as CHIME's detection of ancient hydrogen glow from billions of years ago.
In the pebble-rain picture, the falling pebbles do not simply pile up. As they fall, the iron oxide inside them can react chemically with the hydrogen around them. The reaction strips oxygen from the iron oxide, leaving behind metallic iron — and that freed oxygen combines with hydrogen to make water vapour. In other words, the planet's own building material and its own atmosphere conspire to create water as the planet grows.
The experiment: rock, hydrogen and 1,300 degrees
The experiments were deliberately simple, which is part of their strength. The researchers heated samples of silicate rock — the kind of rocky material that makes up much of Earth's crust and mantle — to temperatures between 1,200 and 1,400 degrees Celsius. Some samples were heated in a stream of hydrogen gas; others, as a control, were heated in helium, an inert gas that does not react chemically. Each run lasted between 15 minutes and three hours.
The difference was striking. In the hydrogen experiments, nearly 80 percent of the iron oxide in the samples was converted into metallic iron, producing water vapour in the process. In the helium experiments, the same reaction did not happen. That control result matters: it shows hydrogen is doing the chemical work, not heat alone.
There was another telling detail. In one hydrogen run at 1,300 degrees, the iron formed rounded droplets — evidence the metal had melted, even though the temperature was below pure iron's melting point; phosphorus in that sample lowered it. In a phosphorus-free sample heated for an hour at 1,250 degrees, the iron instead appeared as solid crystals. Both confirm the same chemistry: iron oxide reduced to iron metal, with water as the by-product.
Lead researcher Zachary Sharp put the scale plainly. The team had hoped to produce iron metal and water, he said, but they had no way of knowing how fast or how much. They were surprised to find that, extrapolated to the early Earth, the process suggests hundreds of oceans' worth of water were likely produced this way. Looking out over a large body of water today, he added, it is hard to imagine that water may have been made 4.5 billion years ago in a hell-like atmosphere — but it very well may have been.
What this does not prove — yet
Before rewriting textbooks, the caveats need to be stated clearly, because they are substantial.
First, this is a laboratory result, not a direct observation of the early Earth. No one can go back 4.5 billion years and watch pebbles fall through a hydrogen atmosphere. The experiments prove the chemistry works under controlled conditions; they do not prove it happened at planetary scale.
Second, the "hundreds of oceans" figure is an extrapolation. It takes a reaction measured in grams of rock over hours and scales it up to a whole planet forming over millions of years. Extrapolations like that are standard in science, but they carry large uncertainties. How long the young Earth kept a hydrogen-rich atmosphere, how much of the falling rock actually passed through the hottest zone, and how much of the water vapour produced escaped back into space are all open questions.
Third, the new mechanism does not have to replace the delivery theory. The most likely truth is a mix: some of Earth's water may have been made in place by pebble-rain chemistry, and some may have arrived later on asteroids and comets. Scientists will test that mix by comparing the chemical fingerprints — in particular, the ratios of hydrogen isotopes — in Earth's water with the fingerprints predicted by each source. If in-place production dominated, the fingerprint should look different from a purely delivered-water fingerprint.
Finally, making water vapour is only half the story. For oceans to exist, the young Earth had to cool enough for that vapour to condense and stay on the surface rather than being lost to space. The early Earth was a violent place — the Moon itself formed from a giant impact, and how the Moon's shrinking faults tell the story of a cooling world shows that these worlds kept changing long after they formed. Water made during formation still had to survive everything that came after.
None of these caveats weaken the result. They define what it is: a proven chemical pathway, with its planetary importance still to be measured. That is how science usually advances — not with one experiment that ends a debate, but with a new mechanism to test against every other line of evidence.
Why it matters beyond Earth
The most exciting implication reaches beyond our own planet. If water can be produced during formation from ordinary rock and hydrogen, then virtually any Earth-sized planet forming in a similar way should end up with water too. Sharp said exactly that: previous ideas relied on an external source, such as comets or meteorites, but this mechanism means almost any Earth-sized planet should have water.
That changes the odds for habitable worlds elsewhere. Astronomers have already found thousands of rocky exoplanets, and one of the biggest questions about each one is whether it has water. If water production is a built-in feature of planet formation rather than a lucky delivery, then watery worlds may be far more common than previously assumed — and so, potentially, are the conditions for life.
It also reframes the search for water on other planets. Beyond asking whether a planet sits in the "habitable zone", researchers may need to ask whether its formation included the right ingredients: iron-bearing rock, a hydrogen-rich early atmosphere, and enough heat. A planet that ticks those boxes may have been born wet, even if it formed in a region once thought too dry.
What scientists will do next
The next steps are clear. Researchers must model the early Earth's atmosphere more precisely — how thick and hot it was, how long it lasted, and how much hydrogen it retained — and test the chemistry across a wider range of rock compositions.
Isotope chemistry will be the decisive test. Different water sources leave different isotopic signatures, and Earth's oceans, mantle and ancient minerals preserve clues about where their water came from. If future measurements show a signature consistent with in-place production, the pebble-rain mechanism will move from plausible to probable.
Timing matters too. Water made during formation would have been present from the very beginning, available for the earliest chemistry leading toward life. Water delivered later would have arrived after the planet cooled. Those two histories imply different conditions for the origin of life.
For now, the study gives us a richer story of our own origins. Earth did not necessarily wait for water to fall from the sky. It may have brewed its oceans from stone and air, in the furnace of its own birth — a planet that made itself wet, and in doing so made life possible.