Editorial NexoScience
What did the universe look like a tiny fraction of a second after it began expanding? Cosmologists have spent decades trying to answer that with a theory called inflation: an almost unimaginably fast stretch of the newborn universe that smoothed it out, flattened it, and left behind a very specific fingerprint in the cosmic microwave background, the faint afterglow of the Big Bang that still bathes the sky today.
That fingerprint has two numbers, and researchers just measured them with more precision than ever. A team combining data from the Planck satellite, the South Pole Telescope, the Atacama Cosmology Telescope, and the BICEP/Keck experiment calculated the spectral index, a number called ns that describes how the universe’s density ripples varied across different scales. The result: ns = 0.9682, give or take 0.0032. They also set an upper limit on r, the tensor-to-scalar ratio, a proxy for gravitational waves generated during inflation that nobody has directly detected yet: r has to be below 0.034.
Those two numbers might sound abstract, but they are precisely what separates one inflation theory from another. Physicists have several competing recipes for what “inflation” actually looked like mechanically, and each one predicts a slightly different combination of ns and r. Two of the most cited recipes are Starobinsky inflation, a model built from a tweak to Einstein’s gravity equations, and Higgs inflation, which tries to recycle the particle that gives other particles their mass into the engine of cosmic expansion.
Adding a fifth data set changed the picture. When the team folded in new galaxy-mapping data from the Dark Energy Spectroscopic Instrument, or DESI, the spectral index shifted upward to ns = 0.9728, give or take 0.0029. That might look like a tiny nudge, but at this level of precision it is enough to matter: the Starobinsky prediction now sits 3.9 standard deviations away from the data, and Higgs inflation sits 2.9 standard deviations away. In physics, that is the kind of gap that makes a once-favorite model start to look shaky.
The data, instead, lines up better with simpler options: so-called monomial potentials, and a family of models called polynomial alpha-attractors. Both describe the inflating universe as rolling down a much plainer slope than the one Starobinsky or Higgs inflation imagine.
The researchers are careful not to oversell the result. The shift only shows up once the DESI data is added, and it stems from a mild, still unexplained disagreement between what the cosmic microwave background says and what the galaxy survey says about the universe’s matter density. That disagreement is real but only marginally significant, and it could fade as more data comes in, exactly the kind of nuance that separates a genuine finding from a headline built on statistical noise. The authors frame this explicitly as the state of the evidence at the end of 2025, not a verdict.
More precise answers are already on the way. Upcoming missions, including the LiteBIRD satellite and the Simons Observatory, are designed to shrink the uncertainty on both numbers by roughly a factor of sixteen within the next decade. At that resolution, a model like Starobinsky inflation would either be confirmed clearly or ruled out for good, and the same goes for its competitors. For now, the universe’s own birth certificate has gotten a little sharper, and one of cosmology’s favorite explanations has some new competition.
Source: L. Balkenhol, E. Camphuis, F. Finelli, et al. “Inflation at the End of 2025: Constraints on r and ns Using the Latest CMB and BAO Data.” The Open Journal of Astrophysics (2026). DOI: 10.33232/001c.164435. arXiv:2512.10613.
