The Hubble Tension: A Crack in the Standard Model of Cosmology
There is a number at the heart of cosmology — a single value that describes how fast the universe is expanding — and right now, two of our best methods for measuring it disagree. Not by a little. Not within the comfortable margins of experimental error. They disagree in a way that has cosmologists genuinely unsettled, and the disagreement refuses to go away.
That number is the Hubble constant, H₀. And the story of its tension may be telling us something profound about the universe we thought we understood.
What the Hubble Constant Actually Measures
Edwin Hubble’s 1929 observation that galaxies are receding from us — and that the farther away they are, the faster they recede — gave us the first empirical handle on an expanding universe. The rate of that expansion, expressed in kilometers per second per megaparsec (km/s/Mpc), is H₀.
For decades, pinning down H₀ was one of cosmology’s great challenges. Estimates ranged from 50 to 100 km/s/Mpc, a factor-of-two uncertainty that made precise cosmological modeling nearly impossible. The Hubble Space Telescope Key Project, completed in 2001, narrowed this to 72 ± 8 km/s/Mpc — a genuine triumph. The era of "precision cosmology" had arrived.
Then we got too precise. And the cracks appeared.
Two Ladders, Two Answers
Today, we measure H₀ in two fundamentally different ways, and they yield stubbornly different answers.
The Late-Universe (Distance Ladder) Method
The first approach builds a cosmic distance ladder from the ground up. It starts with geometric parallax measurements of nearby stars, steps outward using Cepheid variable stars (whose pulsation periods reveal their intrinsic luminosity), and then uses Type Ia supernovae as "standard candles" to reach cosmological distances. From the relationship between distance and recession velocity across thousands of galaxies, you extract H₀.
The SH0ES (Supernova H₀ for the Equation of State) collaboration, led by Adam Riess, has refined this method meticulously. Their 2022 result, published in The Astrophysical Journal Letters, gives H₀ = 73.04 ± 1.04 km/s/Mpc. This is a measurement of the universe as it is today, in the late, evolved cosmos.
The Early-Universe (CMB) Method
The second approach works entirely differently. It looks backward — to the cosmic microwave background (CMB), the afterglow of the Big Bang imprinted on the sky when the universe was just 380,000 years old. The Planck satellite mapped this radiation with extraordinary precision. By fitting the CMB’s temperature fluctuations to the standard Lambda-CDM model of cosmology, Planck doesn’t measure H₀ directly — it infers it. The 2018 Planck results give H₀ = 67.4 ± 0.5 km/s/Mpc.
That is a discrepancy of roughly 5 km/s/Mpc. In statistical terms, the two measurements sit approximately 5 sigma apart — far beyond the 3-sigma threshold typically considered significant in physics. The probability of this being a chance fluctuation is vanishingly small.
Could It Be Systematic Error?
The first, and most scientifically honest, response to any anomaly is to ask: is something wrong with the measurements?
Both camps have asked this question relentlessly. On the distance ladder side, researchers have scrutinized every rung. Could Cepheid measurements be contaminated by crowding in distant galaxies? Could dust extinction be subtly miscalibrated? The SH0ES team has cross-checked using independent distance indicators — the Tip of the Red Giant Branch (TRGB) method, surface brightness fluctuations, water masers — and while some of these yield slightly lower values, none bridge the full gap to Planck’s number.
On the CMB side, the Planck results are extraordinarily robust internally. But they depend on assuming Lambda-CDM is correct. If the standard model is wrong — if there is some physics missing from the early universe — Planck’s inferred H₀ could be systematically off.
The DESI (Dark Energy Spectroscopic Instrument) collaboration, which released its first-year Baryon Acoustic Oscillation (BAO) results in 2024, adds another layer. BAO measurements — using the characteristic scale of galaxy clustering as a "standard ruler" — generally align with the Planck CMB value when interpreted with standard early-universe calibration or combined with other data, reinforcing the early-universe side of the tension. This makes a simple systematic error in Planck less likely.
The tension, in other words, appears to be real.
What Could Explain It?
If the discrepancy is not a measurement error, it is a signal. Here are the leading theoretical proposals, each with its own challenges.
Early Dark Energy
One popular class of models invokes a form of "early dark energy" — a temporary energy component before recombination (the moment the CMB was released) that increases the pre-recombination expansion rate. This would shrink the sound horizon, the characteristic scale imprinted in both the CMB and BAO, and drive the inferred H₀ upward. The idea is elegant, but it tends to worsen other tensions in cosmological data, particularly the S₈ tension — a disagreement between CMB predictions and weak gravitational lensing surveys about the amplitude of matter clustering.
Extra Relativistic Species
Adding extra light particles to the early universe — beyond the three known neutrino flavors — would also alter the sound horizon. Parameterized as N_eff (the effective number of relativistic species), Planck constrains this to 2.99 ± 0.17, consistent with the Standard Model’s prediction of 3.044. There is little room here without violating other constraints.
Interacting Dark Matter or Dark Energy
More exotic proposals include models where dark matter and dark energy interact with each other, or where dark energy is not a cosmological constant (Λ) but a dynamical field that evolves over time. Intriguingly, DESI’s 2024 BAO results showed a mild preference — at roughly 2–3 sigma — for dynamical dark energy over a pure cosmological constant, hinting that Λ in Lambda-CDM may not be the whole story.
Modified Gravity or Late-Universe Physics
Some proposals focus not on the early universe but on modifying how gravity behaves on large scales, or on invoking local inhomogeneities — the idea that we live in a slightly underdense "void" that could bias our local expansion rate measurements upward. Most cosmologists consider this insufficient to explain the full tension.
The S₈ Tension: A Second Crack
The Hubble tension does not stand alone. There is a second, related discrepancy worth naming: the S₈ tension.
S₈ is a parameter that combines the matter density of the universe (Ω_m) with the amplitude of matter fluctuations (σ₈). The CMB predicts a universe that is slightly more clumped — more structured on large scales — than what weak gravitational lensing surveys actually observe. The KiDS-1000 survey, the Dark Energy Survey (DES), and the Hyper Suprime-Cam (HSC) survey all find a lower S₈ than Planck predicts, typically at 2–3 sigma significance.
Two independent tensions, pointing in related directions, both resisting resolution within Lambda-CDM. This is not a coincidence that cosmologists can comfortably dismiss.
What Comes Next
The next few years will be decisive. Several missions are directly targeting these tensions.
Euclid, the European Space Agency’s wide-field space telescope launched in July 2023, will map the geometry and large-scale structure of the universe over a third of the sky, measuring both BAO and weak lensing with unprecedented precision. Its data will sharpen the S₈ measurement and provide independent constraints on H₀.
The Rubin Observatory’s LSST (Legacy Survey of Space and Time), coming fully online in the mid-2020s, will image billions of galaxies and detect hundreds of thousands of Type Ia supernovae, potentially tightening the distance ladder beyond anything previously possible.
LISA (the Laser Interferometer Space Antenna), planned for the 2030s, may offer a completely independent route to H₀ through gravitational wave "standard sirens" — massive black-hole mergers and other low-frequency gravitational-wave sources whose distances can be inferred directly from the gravitational wave signal, bypassing the distance ladder entirely. The first gravitational-wave standard-siren measurement came from the ground-based LIGO/Virgo detection GW170817 in 2017, yielding H₀ = 70⁺¹²₋₈ km/s/Mpc — consistent with both camps, but with too large an uncertainty to adjudicate. More events will narrow this.
Sitting With the Unknown
Here is what I find most intellectually honest to say: we do not know what is causing the Hubble tension.
It may be a subtle systematic error that survives in both camps, waiting for a future instrument to expose it. It may be a genuine signature of new physics — early dark energy, dynamical dark energy, an extra neutrino species, or something not yet written down. It may be telling us that Lambda-CDM, the most successful cosmological model in history, is an approximation to a deeper theory we haven’t found yet.
What I am confident of is this: a 5-sigma tension between two of our most powerful probes of the universe is not noise. It is a message. And the history of physics suggests that messages like this — anomalies that refuse to dissolve — are often the places where the next revolution hides.
The universe has been expanding for 13.8 billion years. We’ve been measuring it seriously for less than a century. The number that describes its expansion rate should, in principle, be one of the simplest things to agree on. The fact that we can’t yet is not a failure of cosmology. It is cosmology doing exactly what science is supposed to do: pressing hard against the edges of what we know, and finding that the edges push back.
Leave a Reply