Skip to content
Saturday, September 26, 202668 days to the Dice Letter centennialNo physics degree or shared belief required.
GOD PLAYS DICE™The magazine of big questions

THE QUESTION LIBRARY ORIGINS

How old is the universe, and how do we know?

About 13.8 billion years, read from the oldest light in the sky and checked against the oldest stars, with one real puzzle still open about how fast space is stretching today.

An oval map of the whole sky on a black background, speckled all over with small blotches of blue and orange, lighter blue and pale yellow, with a large darker blue patch near the center. A color bar at the lower left is labeled Temperature, running from cooler in blue to warmer in red.
The oldest light in the universe, as mapped by ESA’s Planck satellite from its first 15.5 months of data. The warm and cool spots differ by tiny amounts, and their pattern is what lets cosmologists work out the universe’s age.NASA Scientific Visualization Studio (Marit Jentoft-Nilsen, Mark Malanoski), from ESA and Planck Collaboration data, via Wikimedia Commons · Public domainImage source ↗

THE SHORT ANSWER

About 13.8 billion years. The most precise figure, from the European Space Agency’s Planck satellite, is 13.787 billion years, give or take 20 million. Planck read it from the oldest light in the sky, using the standard model of how the universe grows. Other clocks agree. The oldest star clusters in our galaxy come out a little younger, as they must. One puzzle is still open: two ways of measuring how fast space is stretching today disagree.

  • ESA’s Planck satellite gives an age of 13.787 billion years, plus or minus 0.020 billion, if the standard model of the universe is right.
  • A 2020 study of 68 old star clusters in our galaxy found the oldest are about 13.3 billion years old, give or take about half a billion.
  • Hubble dated the cooling remains of dead stars in the star cluster M4 at 12 to 13 billion years, a check that does not depend on the expansion of space.
  • Nearby galaxies say space is stretching at a rate near 73, but Planck’s model predicts about 67 (in the units astronomers use). This gap is the Hubble tension.
  • Webb telescope checks found no simple error in Hubble’s star measurements, but one team using Webb gets about 70, so the cause is still debated.

WHERE THE EVIDENCE STANDS

How sure are we?

  • EstablishedThe universe is about 13.8 billion years old.Planck 2018 final results: 13.787 ± 0.020 billion years (13.797 ± 0.023 from Planck data alone), assuming the standard Lambda-CDM model; NASA’s WMAP found 13.77 billion to within half a percent.
  • EstablishedThe oldest stars and star clusters we can date fit within the universe’s age, once their error bars are counted.Valcin et al. 2020: oldest globular clusters 13.32 ± 0.1 (statistical) ± 0.5 (systematic) billion years; Hubble white dwarfs in M4, 12 to 13 billion years (2002); ESO uranium clock in CS 31082-001, 12.5 ± about 3 billion years (2001); the Methuselah star HD 140283, 14.5 ± 0.8 billion years (2013), overlaps the universe’s age.
  • Supported, still debatedMeasurements of the nearby universe and the early-universe prediction give different expansion rates.SH0ES 2022: 73.04 ± 1.04 km/s/Mpc, a 5-sigma gap from Planck’s 67.4 ± 0.5; the 2026 Local Distance Network: 73.50 ± 0.81; but Freedman’s 2025 result from Hubble and Webb data, 70.4 ± 3%, agrees with Planck within its errors.
  • Not supportedThe Hubble tension or the Webb telescope shows the Big Bang picture is wrong.Both sides of the tension measure an expanding universe and disagree about today’s rate; NASA lists possible causes such as exotic dark energy, exotic dark matter or a change in our understanding of gravity. A 2024 study reported by NASA found the standard model of cosmology likely isn’t broken by early Webb galaxies.
What the labels mean

WHAT WE DON’T KNOW YET

Nobody knows yet why the two expansion measurements disagree. It could be several measurement errors pushing the same way, or a missing piece in the standard model, such as a surprise in dark energy, dark matter or gravity. The 13.8-billion-year figure also assumes that model, though the star clocks, which do not, agree with it.

WHAT WOULD CHANGE THIS ANSWER

Stars that are clearly older than 13.8 billion years, with small error bars, would force a rethink. So would a new model that fixes the Hubble tension and shifts the age. So far, every independent clock points to about the same answer.

About 13.8 billion years. That is how long ago the universe was extremely hot and dense and began the expansion we still see today. The most precise value comes from the European Space Agency’s Planck satellite: 13.787 billion years, give or take about 20 million years.

It is fair to ask how anyone could know that. Nobody was there to start a clock. The answer is that scientists read several different clocks, and they agree.

The first clock is the oldest light in the sky. About 380,000 years after the Big Bang, the universe cooled enough for light to travel freely. That light still fills space as a faint microwave glow. Planck mapped the tiny warm and cool spots in it. The pattern shows what the universe is made of and how it has grown. Scientists then run the standard model of the universe forward to today and read off how much time has passed.

The other clocks are stars. A star cannot be older than the universe it lives in. In one 2020 study, the oldest star clusters in our galaxy came out at about 13.3 billion years. The cooling cores of dead stars, called white dwarfs, gave 12 to 13 billion years in a Hubble study. The first stars formed less than a billion years after the Big Bang, so these ages fit.

There is one real puzzle. Astronomers measure how fast space is stretching today in two ways, and they get different answers. This is called the Hubble tension. It has not changed the measured age, which rests on more than one clock. But it may mean the standard model is missing something. We label the age Established, and the tension Supported, still debated.

THE LONG ANSWER

How does the oldest light tell us the age?

About 380,000 years after the Big Bang, the universe cooled enough for atoms to form, and light could finally travel freely, NASA explains. That light still reaches us from every direction as a faint microwave glow, the cosmic microwave background. It has tiny warm and cool spots, the seeds of later galaxies.

The pattern of those spots depends on what the universe contains and how it has grown. ESA says that by analysing them, scientists can work out the universe’s makeup and history from its birth to today. In practice, they fit a model to the map. The standard model, called Lambda-CDM, describes a universe of ordinary matter, dark matter (unseen matter known from its gravity) and dark energy (whatever is speeding up the expansion), using just six basic numbers. Once those numbers are pinned down, the model can be run forward to the present. The time that has passed is the age.

Planck’s final results were published in 2020. Using Planck’s own data, the age comes out at 13.797 billion years. Adding maps of how galaxies are spread across space gives 13.787 billion years, plus or minus 0.020 billion. Earlier, NASA’s WMAP satellite had found 13.77 billion years, to within half a percent.

This is an inference, not a reading from a cosmic stopwatch. Planck’s measurement of the spot pattern itself depends only weakly on the model. But the paper calls the figures it infers for today’s universe, such as the expansion rate, model-dependent. The star clocks below check the age another way.

Do the oldest stars agree?

They do. Stars are a completely different kind of clock.

Old star clusters. Globular clusters are tight balls of hundreds of thousands to a million stars, among the oldest objects in our galaxy. Stars burn their fuel at rates physics can calculate, so the mix of stars in a cluster reveals its age. In 2020, a team led by David Valcin studied 68 globular clusters. The oldest averaged about 13.3 billion years, give or take about half a billion. After allowing time for the clusters to form, the team estimated the universe’s age at about 13.5 billion years, compatible with Planck.

Cooling embers. A white dwarf is the hot, dense core left behind when a star dies. It no longer burns fuel. It just cools, at a predictable rate. NASA compared this to judging how long ago a campfire burned by measuring the temperature of its coals. The analogy breaks down in one way: nobody can touch these coals, so astronomers rely on models of how white dwarfs cool. In 2002, Hubble found the faintest white dwarfs in the nearby cluster M4 and dated them at 12 to 13 billion years. NASA called this a completely independent reading.

A radioactive clock. In 2001, astronomers using the European Southern Observatory’s Very Large Telescope measured uranium in an ancient star called CS 31082-001. Uranium decays at a known rate, much like the carbon-14 used in archaeology. The uranium clock read 12.5 billion years, with an uncertainty of about 3 billion.

These clocks have caught problems before. NASA recalls that in the late 1970s, age estimates from the expansion ranged from 8 to 18 billion years, and the ages of old stars clashed with the lower values. Better measurements, and the discovery that the expansion is speeding up, settled that clash.

Why do two expansion measurements disagree?

The Hubble constant is how fast space is stretching today. It is given in kilometers per second per megaparsec, where a megaparsec is about 3.26 million light-years, as NASA explains. A value of 70 means a galaxy one megaparsec away moves away at about 70 kilometers per second, and one twice as far moves away twice as fast.

There are two main ways to get it. One predicts today’s rate from the early universe: take Planck’s map, apply the standard model, and calculate. That gives 67.4, plus or minus 0.5. The other measures the nearby universe directly, with a “distance ladder.” Astronomers find distances to pulsing stars called Cepheids, whose rhythm reveals how bright they truly are. They use those to calibrate a kind of exploding star, then measure galaxies farther out. In 2022, the SH0ES team led by Adam Riess reported 73.04, plus or minus 1.04. The team called the gap with Planck a 5-sigma difference, far too large to be a likely statistical fluke, and said its source is still unknown.

Why does this matter for the age? NASA notes that a faster expansion rate leads to a younger age for the universe, and a slower one to an older age. So the tension is really a question about the model that links the early universe to today. ESA offers an analogy: the distance ladder sets an anchor on one shore of a river, Planck sets one on the other, and the stretch in between has not been directly observed. The analogy has a limit. Both anchors are measurements with their own error bars, not fixed posts.

What did the Webb telescope add?

A simple explanation would be a mistake in Hubble’s measurements. Cepheids sit in crowded fields, and their light can blend with neighboring stars. The James Webb Space Telescope sees them more sharply in infrared light. In 2023, Webb checked more than 320 Cepheids and confirmed Hubble’s results. In a 2024 follow-up, it measured about 1,000 Cepheids in five galaxies that hosted eight supernovae, out to 130 million light-years. The team ruled out crowding as the cause at 8.2 sigma.

Not everyone reads the Webb data the same way. Wendy Freedman’s team at the University of Chicago uses different stars, red giants and carbon-rich stars, as distance markers. In 2025, their result was 70.4, plus or minus 3%, which agrees with Planck within the uncertainties. “This new evidence is suggesting that our Standard Model of the universe is holding up,” Freedman said. Riess’s group has replied that the small Webb samples differ from the full Hubble set mainly because of which supernovae they happen to include, and that all Webb methods together give 72.6, plus or minus 2.0.

In April 2026, a large collaboration combined decades of distance measurements into one network. It reported 73.50, plus or minus 0.81, and said this effectively rules out any single overlooked error in local distances. So most local measurements land near 73, one leading team lands near 70, and the cause of the gap is still argued. We label it Supported, still debated.

What do people often get wrong?

“Astronomers found a star older than the universe.” In 2013, a Hubble study dated the nearby star HD 140283, nicknamed the Methuselah star, at 14.5 billion years, plus or minus 0.8 billion. That range reaches down to 13.7 billion, so it overlaps the universe’s age. Earlier estimates had run as high as 16 billion. A better distance measurement shrank the error.

“The Webb telescope broke the Big Bang.” Webb found early galaxies that looked too massive for the standard model. A 2024 study found that growing black holes made some of them look bigger than they are. NASA reports the researchers found no crisis for the standard model. Webb still sees more massive early galaxies than expected, and NASA says one possible reason is that stars formed faster in the early universe.

“The universe is 13.8 billion light-years across.” Space kept stretching while the light traveled. In the current best-fit model, the most distant places we can see are now about 47 billion light-years away, as the UCLA cosmologist Edward L. Wright explains. See How Big Is the Universe, and Does It Have an Edge?

How can you check this yourself?

  • Find the number at its source. Planck’s final paper on cosmological parameters is free on arXiv (1807.06209). Look for the row labeled “Age” in its Table 2.
  • Read the plus-or-minus. A result is a range, not a point. The Methuselah star only looked older than the universe if you ignored its error bars.
  • Do a rough check. The University of Chicago explains that dividing distance by speed, using the Hubble constant, gives a first guess of about 14 billion years. That is close to 13.8, as it should be.
  • Look at the date and the team. The expansion results on this page run from 2022 to 2026 and come from rival groups. A headline saying the tension is “solved” is usually one team’s view.
  • Start at the beginning. For what happened in the universe’s first moments, read What Happened in the First Second After the Big Bang?

After all this detail, the answer is unchanged. The universe is about 13.8 billion years old, and independent clocks agree. The open question is how fast it is expanding today, and why two careful methods disagree.

Deep dive (optional)

Optional detail for the curious. Why is Planck’s age so precise? The key is the typical size of the warm and cool spots on the sky. In the hot early universe, sound waves rippled through the mix of light and gas. The distance they could travel before the light broke free, called the sound horizon, left its mark on the spots. Planck measured how large that scale looks on the sky, the “angular acoustic scale,” to 0.03% precision. It works like a ruler of known length: how big it looks tells you how far away it is. The comparison is loose, because space kept expanding while the light traveled, and the calculation has to include that. A 2001 study led by Lloyd Knox explained why this pins down the age so well: in these models, the age is tightly tied, by a lucky coincidence, to how big that scale looks.

Planck’s final paper (Table 2) lists the age for several data combinations. The temperature map, with a little polarization data added, gives 13.830 ± 0.037 billion years. Data comparing temperature with polarization give 13.761 ± 0.038. All of Planck’s data together, with its map of how gravity bends the microwave light, give 13.797 ± 0.023. Adding baryon acoustic oscillations, a pattern in how galaxies are spread across space, gives 13.787 ± 0.020. These are 68% ranges: the true value should fall inside about two times out of three if the model is right. The same analysis finds that matter, ordinary and dark together, makes up about 31.5% of the universe’s total density today (0.315 ± 0.007) and that the Hubble constant is 67.4 ± 0.5.

The paper is also frank about its limits. In its table of main results, it notes that a total systematic uncertainty of around half of one sigma may be more realistic, and that values should not be over-read beyond that level. That is a small effect on the age, but it is the kind of honesty worth looking for in any measurement.

THREE THINGS TO REMEMBER

  1. The universe is about 13.8 billion years old, and several independent clocks agree.
  2. The oldest star clusters come out a little younger than the universe, which is what should happen.
  3. The Hubble tension is an open puzzle about today’s expansion rate, not a sign the Big Bang is wrong.

WHERE THE AUTHOR’S RESEARCH TOUCHES THIS

Author’s hypothesis. Ricardo Maldonado, author of GOD PLAYS DICE, takes this measured age as his starting point. The book’s glossary dates the hot, dense early state of our universe to about 13.8 billion years ago. Its source notes make the same point this page does: the familiar 13.8 billion years is a model-based inference from Planck’s framework, not a direct reading from a cosmic clock.

His HD-Blast idea, the Author’s hypothesis that our Big Bang may be the aftershock of a higher-dimensional event, asks what may have set off that hot beginning. The book sets a firm condition: any such earlier history must reproduce the measured expansion, the thermal history and the pattern of early ripples, not merely sit alongside them in words. Nothing about the idea has been confirmed.

WORDS WORTH KNOWING

Cosmic microwave background
The faint microwave glow that fills the sky, released about 380,000 years after the Big Bang when the universe became clear enough for light to travel.
Hubble constant
How fast space is stretching today, usually given in kilometers per second per megaparsec.
Megaparsec
A unit of distance equal to about 3.26 million light-years.
Globular cluster
A tight ball of hundreds of thousands to a million stars, among the oldest objects in our galaxy.
White dwarf
The hot, dense core left behind when a star dies. It no longer burns fuel and cools at a predictable rate.
Lambda-CDM
The standard model of cosmology: a universe of ordinary matter, cold dark matter and dark energy (Lambda), described by six basic numbers.

Sources & further reading

  1. Planck 2018 results. VI. Cosmological parameters ↗Abstract: base 6-parameter Lambda-CDM; H0 = 67.4 ± 0.5 km/s/Mpc; Omega_m = 0.315 ± 0.007; angular acoustic scale to 0.03%; results only weakly dependent on the cosmological model. Table 2 (arXiv 1807.06209v4, PDF text extracted): Age 13.830 ± 0.037 (TT+lowE), 13.761 ± 0.038 (TE+lowE), 13.800 ± 0.024 (TT,TE,EE+lowE), 13.797 ± 0.023 (+lensing), 13.787 ± 0.020 (+lensing+BAO) Gyr, 68% limits. Table 1 caption: a total systematic uncertainty of around 0.5 sigma may be more realistic. Context: ESA 2013 first Planck results gave 13.82 billion years (https://www.esa.int/Science_Exploration/Space_Science/Planck/Planck_reveals_an_almost_perfect_Universe); NASA’s WMAP overview says WMAP determined 13.77 billion years to within a half percent (https://science.nasa.gov/mission/wmap/wmap-overview/). Also: NASA Universe overview (https://science.nasa.gov/universe/overview/): around 380,000 years after the big bang, atoms formed and light could travel freely. Knox, Christensen and Skordis, ApJ 563, L95 (2001), https://doi.org/10.1086/338655 (abstract read at arXiv astro-ph/0109232): the CMB age is precise because of a tight (and accidental) correlation of the age with the angle subtended by the sound horizon.
  2. Inferring the age of the universe with globular clusters ↗68 galactic globular clusters; average age of the oldest 13.32 ± 0.1 (stat.) ± 0.5 (sys.) Gyr; inferred age of the universe 13.5 +0.16/−0.14 (stat.) ± 0.5 (sys.) Gyr, compatible with Planck’s 13.8 ± 0.02 Gyr. Abstract read at arXiv 2007.06594.
  3. Hubble Uncovers Oldest ‘Clocks’ in Space to Read Age of Universe ↗White dwarfs in globular cluster M4 (5,600 light-years away, the nearest globular cluster) are 12 to 13 billion years old; first stars formed less than 1 billion years after the big bang; campfire-coals comparison; a completely independent reading not relying on the expansion; late-1970s estimates of 8 to 18 billion years clashed with star ages; globular clusters hold hundreds of thousands to a million stars.
  4. Uranium dates the Universe (eso0106) ↗First uranium measurement outside the solar system, in the old star CS 31082-001, with the VLT and UVES; the uranium clock reads 12.5 billion years; uncertainty about 25%, or about ±3 billion years; compared with carbon-14 dating in archaeology; reported in Nature, February 8, 2001.
  5. Hubble Finds Birth Certificate of Oldest Known Star ↗HD 140283, the Methuselah star, 190.1 light-years away: 14.5 ± 0.8 billion years, overlapping the universe’s age of about 13.8 billion; earlier estimates as high as 16 billion; Hubble parallax five times more precise than Hipparcos. Also opened: NASA, “Webb Finds Early Galaxies Weren’t Too Big for Their Britches After All” (Aug. 26, 2024; https://science.nasa.gov/missions/webb/webb-finds-early-galaxies-werent-too-big-for-their-britches-after-all/): black holes made some early galaxies look more massive; “no crisis” for the standard model; still about twice as many massive early galaxies as expected. Also: Edward L. Wright (UCLA), “Frequently Asked Questions in Cosmology” (https://www.astro.ucla.edu/~wright/cosmology_faq.html, last modified May 24, 2013): distances grow while light travels; the current best-fit model with an accelerating expansion gives a maximum distance we can see of 47 billion light-years.
  6. A Comprehensive Measurement of the Local Value of the Hubble Constant with 1 km/s/Mpc Uncertainty from the Hubble Space Telescope and the SH0ES Team ↗Cepheids in the hosts of 42 Type Ia supernovae; H0 = 73.04 ± 1.04 km/s/Mpc; a 5-sigma difference with Planck plus Lambda-CDM; source of the discrepancy remains unknown (abstract, arXiv 2112.04510). Units: NASA Roman page (https://roman.gsfc.nasa.gov/feature_hubble_constant.html): a megaparsec is about 3.26 million light-years; an object one Mpc away recedes at 70 to 76 km/s, two Mpc at 140 to 152 km/s.
  7. Webb and Hubble telescopes affirm Universe’s expansion rate, puzzle persists (weic2408) ↗Five host galaxies of eight Type Ia supernovae with a total of 1,000 Cepheids, out to NGC 5468 at 130 million light-years; the river-and-two-anchors analogy; 10 to 20 billion year range before Hubble. Paper: Riess et al., ApJL 962, L17 (2024), https://doi.org/10.3847/2041-8213/ad1ddd, rejects unrecognized crowding at 8.2 sigma. NASA Webb blog (Sept. 12, 2023; https://science.nasa.gov/blogs/webb/2023/09/12/webb-confirms-accuracy-of-universes-expansion-rate-measured-by-hubble-deepens-mystery-of-hubble-constant-tension/): more than 320 Cepheids; a faster expansion rate leads to a younger age and a slower one to an older age; possible causes include exotic dark energy, exotic dark matter or a revision of gravity.
  8. New measure of the universe’s expansion suggests resolution of a conflict ↗Freedman’s value 70.4 km/s/Mpc ± 3%, in statistical agreement with the CMB value 67.4 ± 0.7%; red giant and carbon stars; the quoted sentence on the Standard Model holding up. Paper: Freedman et al., ApJ 985, 203 (2025), https://doi.org/10.3847/1538-4357/adce78 (erratum ApJ 993, 252, Nov. 2025: corrects the error bars in Appendix B Figure B1; “do not impact the rest of the paper”). Reply: Riess et al., ApJ 977, 120 (2024), https://doi.org/10.3847/1538-4357/ad8c21: JWST combined three-method result 72.6 ± 2.0; differences explained by sample selection. UChicago explainer (https://news.uchicago.edu/explainer/hubble-constant-explained): speed = distance divided by time gives about 14 billion years.
  9. The Local Distance Network: a community consensus report on the measurement of the Hubble constant at ~1% precision (noirlab2611) ↗H0 Distance Network Collaboration: 73.50 ± 0.81 km/s/Mpc; early-universe predictions closer to 67 or 68; effectively rules out explanations relying on a single overlooked error in local distance measurements; if real, may point to new physics. Paper: Astronomy & Astrophysics 708, A166 (2026), https://doi.org/10.1051/0004-6361/202557993.
  10. GOD PLAYS DICE, Volume OneNotes: the familiar age of roughly 13.8 billion years is a model-based inference from the Planck framework, not a direct reading from a cosmic clock; a higher-dimensional prehistory must reproduce, not merely coexist verbally with, the measured expansion, thermal history and primordial perturbations. Glossary: Big Bang, the hot, dense early state about 13.8 billion years ago. Introduction: the hypothesis that the Big Bang may be the aftershock of a higher-dimensional event.

KEEP ASKING

See a mistake? Report a problem. Corrections are made openly.