FAST and DESI measure a universe still holding atomic hydrogen after 4.5 billion years

2.5 million galaxies, from 0 to 0.41

FAST and DESI measure the cosmic density of neutral atomic hydrogen, Ω_HI, in the redshift window from 0 to 0.41. Hydrogen spectra from the Five-hundred-meter Aperture Spherical Telescope FAST are combined with optical spectroscopy from the Dark Energy Spectroscopic Instrument DESI. The sample covers about 2.5 million galaxies across roughly 12,000 square degrees of sky. The team of Chuan-Peng Zhang, Hong Guo and Yizhou Gu maps that window onto the past 4.5 billion years; Amélie Saintonge is among the authors. The raw decrease is a factor of 1.35 ± 0.10. That tempo is the atomic hydrogen reservoir read across a population, not a snapshot of one galaxy.^[1]

After the conservative systematic corrections from the forward model, the same Ω_HI decline is only a factor of 1.12 ± 0.10. Over those 4.5 billion years the cosmic star formation rate density, CSFRD, fell by a factor of 2.46. The correction does not pull the raw number toward the tempo of star formation; the gap stays open. At fixed stellar mass, the average HI gas fraction evolves by less than 0.2 dex — a base-ten logarithmic unit — across the whole galaxy population. The weak evolution is present across the population, including typical galaxies rather than a few extreme systems. The peer-reviewed paper appeared on 1 September 2026 in Nature Astronomy under the title Weak evolution of cosmic atomic hydrogen over the past 4.5 billion years.^[1]

The size of the reservoir, the tempo of conversion

The team of Zhang, Guo and Gu writes that these quantitative differences rule out rapid depletion of galaxy HI as the primary driver of the late-time CSFRD decline. The atomic hydrogen reservoir thinned by a factor of 1.35 over 4.5 billion years, and by 1.12 after the correction, while star formation fell by a factor of 2.46. The suggested bottleneck sits in the efficiency with which atomic gas is converted into molecular gas; the size of the atomic reservoir does not carry that role. Molecular gas density is known to evolve more closely with star formation; that relation stands here as a known fact, and this paper adds no new molecular-gas census. The paper offers the Ω_HI measurement as a stringent benchmark for models of gas accretion, phase conversion and star-formation regulation.^[1]

The conversion-efficiency reading is a suggestion. This paper does not remeasure molecular gas; it builds an interpretation on the quantitative gap between HI and CSFRD. The DESI DR2 redshift catalogue is not yet public and is subject to the DESI Collaboration data policy, so an independent re-reduction still waits. The weak Ω_HI evolution may also be a residual of the forward model's conservative corrections or of DESI's optical selection. The window from 0 to 0.41 covers 4.5 billion years; the tempo of the atomic reservoir at higher redshift is absent from this study. The stringent benchmark stands for the tempo inside that window.^[1]

The question a quieter sky now carries

What this measurement changes is the public empty-tank story inside a 4.5 billion year window. The universe FAST and DESI measure still holds its atomic hydrogen; what has faded is the tempo of making stars. No species-level destiny follows. The useful question is narrower: which galaxy, under which condition, converts HI into molecular gas, and whose model tests that conversion efficiency. The universal "we" of a fuel-exhausted cosmos splits, in this window, into the modelers' benchmark and the public's bedtime story. Disciplined wonder starts by seeing the gap between a factor of 1.35 and a factor of 2.46.^[1]

If an HI survey outside FAST and DESI covering 0 to 0.41 and a comparable stretch of sky publishes cosmic Ω_HI, this tempo either tightens or stays local. If that survey still leaves the Ω_HI decline over 4.5 billion years far weaker than the CSFRD drop of 2.46, the conversion-bottleneck reading holds; if HI tracks CSFRD closely, the weak evolution from FAST and DESI remains particular to this sample. When you see a faint Milky Way tonight, this window tells of a slowing conversion sitting on an atomic reservoir that is still there. The observable signal is the Ω_HI curve of the next wide HI survey.^[1]