On a mountaintop in Arizona, a four-meter telescope has been doing something that would have been inconceivable a generation ago: conducting simultaneous spectroscopic observations of 5,000 galaxies at a time, night after night, building a three-dimensional map of the universe so vast that it traces the large-scale structure of spacetime across 11 billion years of cosmic history. The Dark Energy Spectroscopic Instrument — DESI — is not a single camera or detector but a robotic system of fiber-optic positioners, each one placing a hair-thin fiber precisely onto the image of a distant galaxy so its light can be split into a spectrum and its redshift measured. Do that to tens of millions of objects and you stop making observations. You start making a census.
The instrument saw first light in 2019 at the Kitt Peak National Observatory, completed a validation survey, and began its main five-year survey in May 2021. By early 2024, the collaboration released results from its first year alone — covering roughly 6 million galaxies and quasars — and those results immediately attracted attention far beyond the usual cosmology community. Not because the map was the biggest, though it was. Because the map was beginning to suggest that something in our standard model of the universe might be wrong.
What dark energy is, and why measuring it is hard
The story begins in 1998, when two independent teams studying Type Ia supernovae discovered that the universe's expansion is not slowing down under gravity's pull, as everyone expected, but accelerating. The only way to make that work mathematically was to invoke a new component of the universe's energy budget — one that permeates all of space, exerts negative pressure, and overwhelms gravity on the largest scales. They called it dark energy, and subsequent measurements suggested it constitutes roughly 68 percent of everything the universe contains.
The simplest description of dark energy is a cosmological constant — a fixed energy density built into the fabric of spacetime, represented by the Greek letter lambda in Einstein's field equations. In this picture, dark energy is constant across time and space. The universe's acceleration rate is fixed. The equation of state parameter w, which describes the ratio of dark energy's pressure to its energy density, equals exactly negative one. That is the prediction. That is what the Lambda-CDM model — the standard model of cosmology — requires.
The problem is that no one has ever measured w directly with enough precision to be certain it stays at negative one across cosmic time. Prior surveys — the Baryon Oscillation Spectroscopic Survey (BOSS), its extension eBOSS, the Planck satellite's measurements of the cosmic microwave background — all gave results consistent with the cosmological constant. But consistent with is not the same as confirmed. The error bars were always large enough to hide something interesting. DESI was built to make those error bars small enough to matter.
Baryon acoustic oscillations as a cosmic ruler
DESI's primary method for constraining dark energy uses a phenomenon called baryon acoustic oscillations, or BAOs. In the early universe, before atoms could even form, the hot plasma of matter and radiation was subject to acoustic waves — pressure-driven sound waves propagating through the medium. When the universe cooled enough for electrons and protons to combine into neutral hydrogen, roughly 380,000 years after the Big Bang, those sound waves froze in place. They left a characteristic imprint in the distribution of matter: a preferred clustering scale of roughly 150 megaparsecs, about 490 million light-years.
That preferred scale is still visible today in the statistical distribution of galaxies. Because we know its physical size from early-universe physics, it acts as a standard ruler — a known yardstick embedded in the universe's structure. By measuring the apparent size of this feature at different redshifts (different epochs in cosmic history), astronomers can track how the universe's expansion rate has changed over time, and from that infer the properties of whatever is driving that expansion. The physics is elegant precisely because it ties the late-universe structure DESI observes to the early-universe physics that Planck measured. Any discrepancy between the two epochs is potentially explosive.
DESI's first-year BAO measurements spanned a redshift range from 0.1 to 4.2 — meaning the survey probes galaxies from relatively nearby, a few hundred million years ago, all the way back to when the universe was barely a billion and a half years old. That range is what makes the results so powerful. Dark energy's influence plays out over cosmic time, and a ruler that works across most of cosmic time is a ruler that can actually catch dark energy changing.
Where the tension lives
When the DESI collaboration combined their BAO measurements with existing data — Planck's CMB observations, supernova surveys including the Dark Energy Survey and Union3 compilations — the preferred value of w shifted away from negative one. In some combinations, the data preferred a model in which dark energy's strength is not constant but evolves with time: stronger in the past, weaker now, or the reverse. The specific parameterization that best fit the data was what cosmologists call w0wa CDM, in which w has an initial value at the present day (w0) and a term describing how it changes with the scale factor of the universe (wa).
The statistical significance of the deviation from the cosmological constant sat at roughly 2.5 to 3.9 sigma depending on which dataset combination was used — not yet at the five-sigma threshold that particle physicists conventionally require before claiming a discovery, but well past the range that can simply be waved away as noise. The collaboration was careful not to overclaim. The results are tantalizing, not definitive. But tantalizing, in cosmology, is the step just before revolutionary.
Several things could explain the tension without overturning the standard model. Systematic errors in any of the combined datasets could shift the inferred value of w. The supernova samples are not perfectly uniform across cosmic time, and there are known concerns about the standardizability of Type Ia supernovae at high redshift. The CMB measurements, while extraordinarily precise, involve their own assumptions about early-universe physics. And DESI itself is still less than halfway through its planned five-year survey. Year two and beyond will either reinforce the signal or erode it.
If it holds, the implications are significant. A time-varying dark energy — called quintessence in one class of theoretical models — would require a scalar field permeating the universe and evolving dynamically, rather than the static cosmological constant. That would be new physics, requiring an extension of the standard model. It would also reopen questions about the ultimate fate of the universe: a cosmological constant implies a fixed future of ever-accelerating expansion ending in a cold, dark, isolated cosmos. A time-varying dark energy might have a different endpoint entirely — or none at all.
DESI is currently mapping galaxies at a rate that has no precedent in observational astronomy. The full five-year survey is expected to catalog roughly 40 million galaxies and quasars. It will be joined by complementary surveys — the Vera C. Rubin Observatory's Legacy Survey of Space and Time, the Euclid space telescope (launched in 2023 and now operating), and eventually the Nancy Grace Roman Space Telescope. Each instrument approaches the dark energy problem from a different angle: weak gravitational lensing, galaxy clustering, supernova rates, void statistics. What DESI is beginning to find, the others will either confirm or contradict.
That convergence of evidence — or divergence from it — is likely to be the defining cosmological question of this decade. For now, the map keeps growing, one spectrum at a time, in the dark above the Arizona desert.
