Sven Erik Matzen

Software Architect | Cloud & Security Expert | Scalable Solutions

The Cosmic Tension: Why the Universe Seems to Have Two Rates of Expansion

EU label: fully AI-generated content Fully AI-generated content. This article was generated using AI systems (e.g. Claude Code, Perplexity) and may contain errors or hallucinations.

← All articles

Astrophysics · 2026-06-20 · approx. 30 minutes

The Hook: A Single Number That Refuses to Add Up

Imagine you measure the length of a table with a tape measure and get 180 centimeters. Then a colleague measures the same table with a laser rangefinder and gets 165 centimeters. You repeat both measurements, refine the instruments over years, rule out every conceivable source of error — and the difference remains. It even grows more stable. At some point it becomes clear: it is not the instruments. Either the table is not a rigid thing, or your understanding of what "length" even means is incomplete.

The science of cosmology has been in exactly this situation for a good decade. The measured quantity is called the Hubble constant, H₀ for short, and it describes how fast the universe is currently expanding. There are two fundamentally different, each highly precise methods for determining it. One persistently yields a value of about 67.4 kilometers per second per megaparsec. The other just as persistently yields about 73. The two numbers should be identical — it is the same physical quantity of the same universe. But they are not, and the discrepancy can no longer be dismissed as bad luck, sloppiness, or a statistical outlier.

This conflict goes by the name Hubble tension. It is today probably the most prominent open question in cosmology and is regarded by many as a possible crack in the foundation of our standard model of the universe. The statistical significance of the deviation now stands at about 5 to 6 standard deviations — a level that in physics normally counts as a "discovery." Only, no one yet knows what has been discovered: a hidden measurement error, or new physics beyond all previous theories.

This article takes you from Edwin Hubble's first, grossly wrong measurement, through the decades-long "factor-of-two war" between two astronomers, to the most recent data from 2024 to 2026, in which the James Webb Space Telescope, surveyed gravitational lenses, and the DESI survey have not resolved the tension but sharpened it.

Part 1: What the Hubble Constant Actually Describes

A Universe That Stretches

In 1929 the astronomer Edwin Hubble published an observation that shifted the worldview: the farther a galaxy is from us, the faster it moves away from us — and in an almost linear relationship. Double the distance means double the recession velocity. This proportionality is the Hubble-Lemaître law, and the proportionality factor between distance and velocity is precisely the Hubble constant.

The right mental picture is important: the galaxies are not flying through a fixed space like fragments of an explosion. Rather, space itself expands, and the galaxies are, as it were, carried along. The classic image is rising raisin bread: as the dough swells, every raisin moves away from every other. Seen from any single raisin, the more distant raisins recede faster — exactly the Hubble behavior. There is no center; every observer sees the same thing.

The mathematical groundwork had been laid two years earlier, in 1927, by the Belgian priest and physicist Georges Lemaître, who derived an expanding universe from Einstein's general theory of relativity and predicted the recession of the galaxies. That is why the law today rightly bears both names.

What the Unit "km/s/Mpc" Means

The Hubble constant is given in the somewhat unwieldy unit kilometers per second per megaparsec. One megaparsec (Mpc) is about 3.26 million light-years. The statement "H₀ = 70 km/s/Mpc" means, in plain terms: a galaxy one megaparsec away recedes from us at 70 km/s; a galaxy at two megaparsecs at 140 km/s, and so on.

H₀ has a second, deeper meaning as well. Take its reciprocal and you get a time — the so-called Hubble time, a rough estimate of the age of the universe. A larger H₀ means a faster expansion and thus a younger universe; a smaller H₀ an older one. The Hubble tension is therefore not merely academic number-crunching: behind the values 67 and 73 stand slightly different stories about the age, size, and composition of the cosmos.

Part 2: The Historical Run-Up — From 500 to 70

Hubble's Spectacularly Wrong Starting Value

It belongs to the humility of this science that Hubble's first value for "his" constant was grossly off: he estimated H₀ at about 500 km/s/Mpc, later raised slightly to 530. That is roughly seven times the value accepted today. The reason was not a thinking error but flawed distance measurements: Hubble massively underestimated the distances to the galaxies, because the calibration objects available at the time were poorly calibrated. Too small a denominator (distance) at the same velocity yields too large an H₀.

This episode is instructive: the Hubble constant stands or falls with the accuracy with which we determine distances in space. And distances in the universe are notoriously hard to measure, because we cannot travel there and check. We see only brightnesses in the sky — and must infer distances from them.

The Factor-of-Two War

In the second half of the 20th century, H₀ shrank continuously in the estimates, yet a stubborn dispute solidified. Two camps, led by two great astronomers, fought a decades-long duel:

  • Allan Sandage (a student of Hubble) and his collaborators favored a low value around 50 km/s/Mpc.
  • Gérard de Vaucouleurs and his camp arrived at a high value around 100 km/s/Mpc.

The two values differed by a factor of two — and correspondingly people argued over a universe that was either about 10 or about 20 billion years old. This "factor-of-two war" shaped cosmology for decades and was often bitter. It ended only when a new instrument placed distance measurement on a more solid basis.

The Hubble Telescope Brings Order — Provisionally

A central mission of the Hubble Space Telescope, launched in 1990, was precisely to settle this question. The "HST Key Project," led by Wendy Freedman, measured the distances to numerous galaxies using Cepheid stars (more on these shortly) more precisely than ever before. Around the turn of the millennium, the values converged on about 72 ± 8 km/s/Mpc — a compromise that, in a sense, met both old camps in the middle and ended the factor-of-two war.

It seemed the Hubble constant was thereby essentially settled. In fact, however, the real, far subtler story was only beginning — because in parallel, a completely different method for determining H₀ was maturing, one with no need to look at nearby galaxies at all.

Part 3: Two Roads to the Same Number

Today's tension arises because two methodologically independent procedures arrive at two different values. They are often called the "late" (local) and the "early" (cosmological) methods. They could hardly be more different — and that is exactly what makes the conflict so explosive.

Method A: The Cosmic Distance Ladder (the "Late" Universe)

The first method measures the expansion directly in the here-and-now observable, comparatively nearby universe. It is called the cosmic distance ladder, because it consists of successively built "rungs," each calibrating the next:

  1. First rung — geometric calibration: For the nearest stars, the distance can be determined purely geometrically via parallax (the apparent shift of a star against the background as the Earth orbits the Sun). This is the only rung without any physical assumption — pure trigonometry.
  2. Second rung — Cepheids: Cepheids are pulsating stars with a remarkable property, discovered in 1908 by Henrietta Leavitt: the duration of their brightness variation depends directly on their true luminosity (the period-luminosity relation or Leavitt law). Measure how long a Cepheid takes for one pulsation, and you know its actual brightness — and comparing it with the apparent brightness in the sky yields the distance. Cepheids reach far enough to survey galaxies at medium distance.
  3. Third rung — Type Ia supernovae: A Type Ia supernova is the thermonuclear explosion of a white dwarf and reaches a very well-known peak brightness — it is a "standard candle." Because such supernovae are extremely bright, they can be seen across billions of light-years. In nearby galaxies whose distance is already known via Cepheids, the supernovae are calibrated; then they are used to push into the depths of the cosmos.

The team that operates this ladder most accurately today is called SH0ES (led by Adam Riess, who shared the 2011 Nobel Prize in Physics for the discovery of the accelerating expansion). Its current value: H₀ ≈ 73 km/s/Mpc, with an uncertainty of about one percent.

Method B: The View into the Early Universe (the "Early" Method)

The second method needs no distance ladder at all. It reads H₀ off the cosmic microwave background — that faint radiation the universe emitted about 380,000 years after the Big Bang, when it first became transparent. This "baby photo" of the cosmos contains a fine pattern of tiny temperature fluctuations, which the ESA Planck mission mapped with high precision.

From this pattern — with the help of the cosmological standard model, called ΛCDM (Lambda-CDM, for the dark energy Λ and cold dark matter) — the entire later development of the universe can be computed forward, including its present rate of expansion. This is therefore not a direct measurement of today's expansion, but a model-based prediction: "If ΛCDM is correct and the early universe looked like this, then H₀ today must have this value." Planck's result: H₀ = 67.4 ± 0.5 km/s/Mpc.

The Confrontation

Here the two worlds collide:

Method Procedure H₀ value (km/s/Mpc) Character
SH0ES (local/"late") distance ladder: Cepheids + Type Ia supernovae ≈ 73.0 (± ~0.9) direct measurement of today's universe
Planck (cosmic/"early") microwave background + ΛCDM model 67.4 (± 0.5) model-based prediction from the Big Bang echo

Both camps have refined their procedures over years, and both have shrunk their error bars so far that the two values no longer overlap today. The difference amounts to about 9 percent — small enough that it once vanished into the noise, but now secured at more than five standard deviations. This, precisely, is the Hubble tension.

Part 4: The Decisive Question — Measurement Error or New Physics?

When two measurements stubbornly fail to agree, there are basically only three possibilities. First: the "late" method has a hidden systematic error. Second: the "early" method has one — or, more precisely, the ΛCDM model with which it computes is incomplete. Third: both measure correctly, and the universe actually behaves differently than we think. Research over the past years has worked systematically through these possibilities.

The Prime Suspect: Is the Error in the Distance Ladder?

For a long time, the most obvious suspicion was that the Cepheid rung of the distance ladder has a hidden problem. Cepheids often lie in dense, dusty regions of their galaxies. Two effects could falsify their brightness: dust, which dims the light and makes it appear redder, and crowding, that is, neighboring stars that merge with the Cepheid in the telescope and make it appear brighter than it is. Both would distort the distances and could in theory explain the entire "73" measurement.

This is where the James Webb Space Telescope (JWST), in operation since 2022, came into play. JWST sees in infrared light far more sharply than the old Hubble telescope and penetrates dust much better. It was thus the ideal tool to test exactly these suspects: are Hubble's Cepheid measurements falsified by dust and crowding?

Webb's Surprising Answer: The Ladder Holds

The answer that crystallized in the years 2023 to 2025 was, for many, sobering: JWST confirmed the Hubble measurements. The sharper Webb images showed that the old telescope's Cepheid distances were not systematically falsified by dust or crowding. The local value of about 73 stood firm. Put bluntly: humanity's best telescope was sent out to find the suspected error — and found none.

With that, the burden of proof shifted. JWST has not resolved the tension but deepened it. If the local measurement is robust, then the problem lies either in the "early" method — or in the physics in between.

Independent Cross-Checks

A finding becomes especially convincing when several mutually independent methods support it. This is increasingly the case for the high local value:

  • Gravitational lenses (TDCOSMO, 2025): When the light of a distant quasar passes a massive foreground galaxy, it is bent and reaches us along several paths of different length. From the minimal time delays between these images, H₀ can be determined — entirely without a distance ladder. A "blind" 2025 analysis with eight such lens systems and improved velocity measurements (using JWST, Keck, and VLT) yielded H₀ ≈ 72 km/s/Mpc — quite compatible with the ladder.
  • Tip of the red-giant branch (TRGB): An alternative calibration of the supernovae uses not Cepheids but the maximum brightness of old red giant stars as a standard candle. This method, championed by Wendy Freedman, yields values between about 69 and 72 — tending somewhat lower than the Cepheids, but still clearly above the Planck value of 67. So it does not exactly bridge the gap.
  • Type II supernovae: A method that measures Type II supernovae directly via the physical modeling of their spectra — entirely without a rung ladder — likewise yields a value near the local scale.

The overall picture is remarkable: almost all "late" methods land above 70, almost all "early" ones land below it. The dividing line does not run between individual instruments but between local measurements and model-based predictions from the early universe. This increasingly suggests that no single instrument is to blame, but rather that our model of the universe has gaps in one place.

Part 5: The Attempted Solutions — and Why None Fully Convinces

If the measurements are correct, the physics must be amended. Theoretical creativity has accordingly exploded: one review classified the proposed explanations into 11 major categories with over 120 subcategories — a measure of how open the field is and how much it is searching for an idea that fits everything together. Three thrusts are especially influential.

Early Dark Energy

The currently most discussed candidate is Early Dark Energy (EDE). The idea: in the very young universe, shortly before the formation of the microwave background, there existed for a short time an additional, repulsive form of energy (typically a scalar field), which then rapidly disappeared again. This brief "energy boost" would shrink a certain reference length in the early universe — the so-called sound horizon. Since the Planck method uses precisely this length as a yardstick to infer H₀, a smaller sound horizon would shift the "early" prediction upward, closer to 73.

EDE is elegant because it intervenes at exactly the right spot. But it has side effects: it also alters other quantities of the cosmic pattern, and these alterations do not fit all the data perfectly. The current state is that EDE can mitigate, but not entirely and consistently eliminate, the tension. The data remain too ambiguous to confirm or clearly reject EDE — only the next generation of experiments is likely to decide here.

Evolving Dark Energy and the DESI Finding

A second trail, much noted from 2024 to 2026, comes from DESI (Dark Energy Spectroscopic Instrument), which measures the three-dimensional distribution of millions of galaxies and from it reconstructs the expansion history via the so-called baryon acoustic oscillations (a "standard ruler" in space). The DESI data contain an explosive hint: the dark energy might not be constant but may change over cosmic time — a finding in tension with the simple ΛCDM model, in which Λ (the dark energy) is an unchanging constant.

DESI itself, combined with background data, arrives at an H₀ around 68.5 km/s/Mpc, placing it close to Planck. More exciting, though, is another analysis: if you evaluate H₀ separately for different distance ranges (redshifts), a declining trend appears — H₀ appears higher in the nearby universe than in the distant one. Some researchers interpret this as a possible way to resolve the tension "naturally"; others caution that such trends can depend on the chosen analysis and are not yet secured. Caution is warranted here: in my view, these DESI hints are highly interesting but not yet robust enough to count as a solution — the field is discussing them actively and controversially.

Are We in an Unusual Place? The Local Underdensity

A third, comparatively conservative idea forgoes new physics: perhaps our cosmic neighborhood happens to lie in a below-average-density bubble of the universe. In such a "local void," matter would be drawn outward, and the locally measured expansion would appear faster than it is on the cosmic average. This would explain the high local value without touching ΛCDM. However, such an underdensity would have to be quite large and pronounced, and the surveys so far of the matter distribution in our surroundings provide no convincing evidence for it. Most experts consider this explanation insufficient on its own.

The Field's Honest Verdict

As of 2026 there is no generally accepted solution. Every proposal solves one problem and creates a new one elsewhere. This is not a sign of stagnation but of a genuine scientific crisis in the constructive sense: a standard model successful for over two decades is hitting a limit, and no one yet knows whether it needs a repair patch or a deeper upheaval. Historically, such moments in physics have often been the harbingers of great breakthroughs.

Part 6: Why This Is More Than a Quarrel over Decimal Places

It Is About the Ingredients of the Cosmos

The Hubble constant is interwoven with almost all other cosmological quantities — with the age of the universe, with the amount of dark matter and dark energy, with the history of structure formation. A genuine, physical discrepancy in H₀ would mean that our inventory of the universe is incomplete in at least one place. Since about 95 percent of the cosmic energy content already falls to the puzzling components of dark matter and dark energy, which we do not directly understand, a new building block — such as Early Dark Energy or an evolving dark energy — would be an enormous gain in knowledge.

A Lesson in Precision and Accuracy

The Hubble tension is also a lesson in the difference between precision (low random scatter, small error bars) and accuracy (hitting the true value). Both camps have become extremely precise — their error bars are tiny. But it was precisely this precision that made the tension visible in the first place: as long as the error bars were large, 67 and 73 overlapped easily. Only the shrinking of the uncertainties made the conflict inescapable. This is a paradoxical but recurring experience of science: better measurements do not always resolve old questions — sometimes they first bring the truly deep problems to light.

The Methodological Parallel

It is worth looking at the structure of this dispute, for it resembles other great turning points. In the history of quantum entanglement (see Spooky Action at a Distance: Quantum Entanglement from Einstein to the Quantum Internet), a fundamental dispute remained undecidable for decades until a precisely posed question — Bell's inequality — made it experimentally testable. And in bioenergetics (see The Molecular Turbine: ATP Synthase and the Engine of Life), Peter Mitchell's initially ridiculed theory was confirmed not by authority but because it made testable predictions that held. The Hubble tension stands today at exactly this point: a clearly measurable discrepancy presses for a decision, and it is the quality of the data — not the volume of opinions — that will ultimately tip the scales. The mechanism of scientific progress is the same in all three cases.

What Comes Next

The decision will not come from a single measurement but from the convergence of many. Several lines of development are taking shape. Gravitational waves from merging neutron stars provide "standard sirens," whose distance can be determined entirely independently of light-based standard candles — a wholly new, third pillar. The ongoing DESI and upcoming surveys will resolve the history of dark energy more sharply. And new ground- and space-based telescopes will further improve both the microwave background and the local distance ladder. Should all independent roads continue to yield two separate values, the case for new physics would be hard to dispute.

The Central Takeaway

The Hubble tension condenses a great idea into a single, stubborn number:

  1. Empirically there are two independent, highly precise roads to determining the universe's rate of expansion — the local distance ladder (≈ 73) and the model-based prediction from the Big Bang echo (≈ 67). They do not agree, and the difference is secured at more than five standard deviations.
  2. Methodologically the James Webb Telescope cleared away the most obvious suspect — a measurement error in the Cepheid ladder — and thereby did not defuse the tension but sharpened it.
  3. Theoretically the standard model ΛCDM is thus under pressure, without any replacement (Early Dark Energy, evolving dark energy, local underdensity) so far being fully convincing.

A concrete prompt to act on: The next time you see two "hard numbers" set against each other in a discussion — at work, say, two metrics that should measure the same thing but diverge — do not reflexively treat the difference as a measurement error to be argued away. Instead ask systematically: are both procedures really independent? Does one side measure directly, the other via a model? And what if both numbers are correct and a hidden assumption is false instead? Exactly this attitude — taking the discrepancy seriously rather than smoothing it over — is driving cosmology right now to its most exciting frontier.

A reflection question: The Hubble tension became visible only when both measurement methods grew more precise — better data here solved no problem but first uncovered a deeper one. Where in your own models — professional or organizational — might it be that not more accuracy, but a hitherto unquestioned basic assumption, is the real bottleneck?

Cross-References in the Vault

  • Spooky Action at a Distance: Quantum Entanglement from Einstein to the Quantum Internet – As with the Hubble tension, here too a long-standing fundamental dispute became decidable only through a precise, experimentally testable question; moreover, the cosmic microwave background is itself a window into the quantum-physical beginnings of the universe.
  • The Molecular Turbine: ATP Synthase and the Engine of Life – There too, an initially ridiculed theory (Mitchell's chemiosmosis) prevailed not through authority but through testable predictions — the same logic of knowledge that will ultimately decide the Hubble tension.
  • Der Wettlauf mit der (um die) KI – Both fields live on ever larger volumes of data and ever more precise models; the Hubble tension shows exemplarily that more precision alone guarantees no truth.

Sources and Further Reading

Created as part of the daily learning workflow. Field of interest: Astrophysics. Estimated reading time: ~30 minutes.

← All articles