FAST Telescope: Fewer Stars, Hydrogen Nearly Unchanged
Measurements from FAST and DESI across 2.5 million galaxies show hydrogen barely changed in 4.5 billion years, while star formation fell by half. Published in Nature Astronomy.
A team led by the Chinese Academy of Sciences, working with the DESI (Dark Energy Spectroscopic Instrument) project, has published a measurement that upends a common explanation for why the universe is making stars at a slower and slower pace. The research shows that the star formation rate has dropped to less than half its former value over the past 4.5 billion years, while the supply of neutral atomic hydrogen, the gas from which stars indirectly form, has declined only slightly. The peer-reviewed paper appeared in Nature Astronomy on September 1, 2026, under a title describing the weak evolution of cosmic atomic hydrogen over the past 4.5 billion years, with the identifier 10.1038/s41550-026-02965-9.
Why the universe finds it increasingly hard to produce new stars has long sat at the center of research into how galaxies form and evolve. The standard assumption held that cold gas, the fuel for star birth, is gradually used up as the universe ages, and that this alone would drive down the star formation rate. If that assumption were correct, the decline in star formation should track a similarly steep, synchronized drop in the supply of cold gas. The new research finds no such synchrony. Neutral atomic hydrogen, denoted HI, is the crucial link between the large-scale circulation of cosmic gas and star formation inside galaxies, and it is detected through an extremely faint radio line at a wavelength of 21 centimeters. The signal from any single distant galaxy is usually too weak to pull out of the noise, which has left astronomers choosing between deep observations that cover only a sliver of the sky and wide surveys that lack the needed sensitivity.
To get around that trade-off, the researchers combined the exceptional radio sensitivity of China’s FAST telescope (Five-hundred-meter Aperture Spherical Telescope), a 500-meter dish built inside a karst sinkhole, with the massive optical spectroscopic survey run by DESI. The team analyzed a sample of roughly 2.5 million galaxies spread across about 12,000 square degrees of sky, nearly a third of the entire celestial sphere. They used a technique called hydrogen spectral stacking: the faint radio signals of individual galaxies, each undetectable on its own, are aligned according to the galaxies’ precisely measured redshifts and then added together, pulling an average signal out of the noise with far greater statistical precision than earlier methods allowed.
The measurements revealed a divergence. The star formation rate 4.5 billion years ago was about 2.5 times higher than it is today, or more precisely a factor of 2.46 according to the paper itself. The density of neutral atomic hydrogen over the same period, according to the accompanying announcement, was only about 1.4 times higher than today. The paper offers a sharper picture here: the raw measured difference in hydrogen density comes out to a factor of 1.35, with an uncertainty of 0.10, and after conservative corrections that value drops to 1.12, again with an uncertainty of 0.10. In other words, the press announcement and the paper do not fully agree on the rounded figure, but both point to the same pattern: hydrogen was not disappearing anywhere near as fast as star formation was declining.
According to the authors, the finding shifts the research question away from whether the gas is disappearing and toward why it is becoming harder to form stars despite still-abundant supplies of neutral hydrogen. Stars form primarily in denser clouds of molecular gas, while neutral atomic hydrogen occupies an intermediate step between the large cosmic gas reservoir and the formation of that molecular gas. The researchers believe the most significant recent changes are not happening in the total HI supply, but in how gas moves through the so-called baryon cycle: as the inflow of gas from the cosmic web weakens and gas density drops, the efficiency of converting atomic hydrogen into molecular gas likely declines, even as the overall HI supply stays relatively stable.
The study was led by a team from the National Astronomical Observatories of China, the Shanghai Astronomical Observatory of the Chinese Academy of Sciences, and Shanghai Jiao Tong University, with collaborators from the DESI collaboration and research institutions across Asia, North America, and Europe. The paper’s first author is Chuan-Peng Zhang, according to the full author list given on the paper’s preprint page (arXiv 2607.05326). An earlier published transliteration of the name as Cheng-Peng Zhang was incorrect and is corrected here.
The study itself does not prove the mechanism behind the declining efficiency of converting hydrogen into molecular gas, but points to it indirectly through the divergence between the two curves. The spectral stacking method produces an average value for a large sample of galaxies and says nothing about individual cases, or about why any one particular galaxy is slowing down or speeding up its star formation. The paper also does not directly measure molecular hydrogen; its conclusion about that gas declining is instead inferred from the gap between the star formation curve and the atomic hydrogen curve. The authors themselves describe the finding as a new foundation for further measurements of cosmic gas circulation, not as a final explanation for why the number of new stars being born in the universe is falling.
