For 20 years, astronomers studying how the first galaxies formed used a model that turned out to be wrong for nearly half of them, as Universe Today reported when the findings were published. A new survey has found that hydrogen gas clouds — the fuel that feeds star formation — surround early galaxies in ways the field never systematically accounted for, according to a peer-reviewed study published in the Astrophysical Journal in March 2026.
The clouds are called hydrogen halos, or by their technical name, Lyman-alpha nebulae. They are the raw material from which stars form: a galaxy's hydrogen supply, extending well beyond the bright core that telescopes normally see. Understanding how these halos behave matters because they determine whether a galaxy can keep making stars, and for how long.
The Hobby-Eberly Telescope Dark Energy Experiment (HETDEX), a survey run from the McDonald Observatory in Texas, has now compiled the largest catalog of these objects ever assembled. Using 156 spectrographs — tools that spread light into its component colors, revealing the composition and motion of distant gas — HETDEX surveyed over 540 square degrees of sky, an area covering roughly 2,000 full Moons. The project generated nearly half a petabyte of data across 1 million light sources dating back 10 to 12 billion years, to an era called Cosmic Noon, when the universe was roughly 2 to 3 billion years old.
The result: 33,612 hydrogen halos confirmed, according to the team's analysis, compared to roughly 3,000 known before. That is a tenfold increase in the sample size, as HETDEX's own summary notes, from a handful to a statistical catalogue.
But the more significant finding is what those 33,000 objects revealed about the models. Forty-seven percent of them — nearly half — show extended hydrogen emission that fits a two-component model, with a compact core surrounded by a diffuse outer envelope. The field, for two decades, had been analyzing these objects using a single-component model that captures only the core. The extended component was there all along; it simply fell below the detection threshold of earlier instruments.
"The faintest systems simply aren't bright enough to fully reveal how large they are," the researchers noted. With deeper observations, the authors suggest, essentially every source in the catalogue would likely show the same two-component structure. The 47.5 percent is probably a floor, not a ceiling.
The implication cuts into the theoretical foundation. Galaxy formation models have depended on the single-component framework to simulate how gas flows into early galaxies and sustains star formation. If the compact core is only half the picture, and the extended envelope is a universal feature that earlier surveys simply couldn't see, then the models need to account for a mode of gas delivery that has been effectively invisible until now.
In other words: the model wasn't wrong in a minor way. It was missing half the geometry of how early galaxies were fed.
Whether the correction ripples outward to other predictions — star formation rates at Cosmic Noon, the mass of hydrogen reservoirs, the timeline for when galaxies exhaust their fuel — is what the field will now test. HETDEX is ongoing, and the team is working to push sensitivity deeper in follow-up observations.
The short version: astronomers had been studying the same 3,000 objects for 20 years. Now they have 33,000, and what those 33,000 show is that the model needs a rewrite.