The Biggest Things That Exist, With the Error Bars Left On

This is Vishala Space. Every video about the biggest star says the same thing. Put it where the Sun is, and it swallows Saturn. So here it is, at the Sun's place: one thousand eight hundred and forty times the Sun's radius, its best published size. And here is Saturn's orbit, nine point five seven times farther out than Earth. The star stops short of it. Only the top of its error bar gets there.

Tonight, every size comes with its error bar left on. And the biggest things in the universe start to look different. A star's size is one of two sums. Measure how wide it looks in the sky, and multiply by how far away it is. Or measure how bright it is and how hot, and work out how big a surface must be to shine that much. Both sums need the distance. Get the distance twenty percent wrong, and the size is twenty percent wrong.

Take UY Scuti, the biggest star on every list for a decade. In 2013, a telescope array measured its disc at five point four eight thousandths of an arcsecond, with an error under two percent. Then came a distance from 1970: about nine thousand five hundred light-years, give or take ten percent. The answer: one thousand seven hundred and eight times the Sun's radius, plus or minus one hundred and ninety-two. Almost all of that error bar was the map, not the telescope. Newer distances put the star at about five thousand nine hundred light-years. Same disc, new distance: about a thousand Suns.

The brightness route says about nine hundred. UY Scuti did not shrink. The map got better. And at nine hundred, it would not even reach Jupiter. So let us climb, with the bars on. Betelgeuse, the red shoulder of Orion, the one giant you can find with your own eyes. Seven hundred and sixty-four Sun radii, plus one hundred and sixteen, minus sixty-two.

In the Sun's place it would reach past Mars and the asteroid belt, but not Jupiter. Two papers, two years apart, put it at five hundred and fifty and seven hundred and twenty light-years away. Thirty percent apart, for the nearest red supergiant we have. Why not just measure the parallax? Because a star like this is six times wider than Earth's whole orbit, its surface is a handful of boiling cells, and the bright spot a telescope locks onto drifts. A negative distance.

The maths is telling you it cannot find the centre. AH Scorpii: one thousand four hundred and eleven in 2013; re-measured in 2024 at about twelve hundred and thirty. Every rung of this ladder is soft. The higher you climb, the softer it gets. Here is the swap. First, Stephenson 2-18, the top of every list. Nobody has measured its disc.

Its size is the brightness route: a 2026 measurement of its light, about three hundred and twenty thousand Suns, give or take a hundred and fifteen thousand, at an assumed surface of about five thousand three hundred degrees Fahrenheit. That gives one thousand eight hundred and forty Sun radii, plus three hundred and ten, minus three hundred and sixty. The distance is a cluster distance, and three teams doubt the star is even in the cluster. Now VY Canis Majoris. Its disc was measured directly, and its distance comes from radio parallax, good to ten percent. One thousand four hundred and twenty, plus or minus one hundred and twenty. The bottom of Stephenson's bar sits inside the top of VY's.

The biggest star anyone has measured well is about fourteen hundred times the Sun's radius, give or take a hundred and twenty. The ones listed above it are the ones we trust least. And there is a ceiling. No cool supergiant in any galaxy we have checked shines above about three hundred and twenty thousand Suns. At their temperature, that caps a red supergiant near fifteen hundred and thirty. The famous two thousand one hundred and fifty needs a star above that ceiling and colder than the cold limit. Why is the bar so wide?

Three inputs, each soft. How far. How bright. How hot. Distance goes straight in. Temperature goes in squared: a slip of one hundred degrees on a star this cool moves the size six percent. And the edge itself is soft.

Push into the surface. This is not a textured ball. The best images of Betelgeuse show a handful of convection cells, each larger than Earth's whole orbit, rising and sinking over months. In 2019, one cool patch and a dust cloud made half of Betelgeuse ten times dimmer. Where does a star like that end? The number we quote is where a model says the light comes from. Still, the scale is real.

Light crosses VY Canis Majoris in about one hour and fifty minutes. It crosses the Sun in under five seconds. A passenger jet flying around VY at five hundred and sixty miles an hour would need about eight hundred years. Now pull back. The star becomes a point. A black hole's size is its mass times a constant. Every Sun's worth of mass adds about one point eight miles to its radius.

So every error bar on a mass is an error bar on a size, one for one. Start at home. Sagittarius A star, the black hole at the centre of our galaxy: four point three million Suns, known to about one percent from the orbits of single stars. Its horizon is about sixteen million miles across. The dark shadow around it, about forty-one million miles, a little wider than Mercury's orbit. Then M87 star, fifty-five million light-years away. Six and a half billion Suns, plus or minus seven hundred million.

Its horizon alone is two hundred and fifty-seven times the Earth-Sun distance across. Wider than the distance from the Sun to Voyager 1, the farthest thing we have ever built. Light takes a day and a half to cross it. Now the heavyweights. Holmberg 15A, the giant galaxy at the heart of a cluster about seven hundred million light-years away. In 2019, a team modelled the orbits of its stars and found forty billion Suns, plus or minus eight. In 2025, another team, with a different telescope and model, found twenty-two billion, plus two, minus two.

The error bar here is the model. So the heaviest black hole anyone has actually weighed is somewhere between about twenty and forty billion Suns. Even the small end swallows the whole solar system: a horizon eight hundred and fifty times the Earth-Sun distance across. Then TON 618, the quasar on every list. Sixty-six billion Suns, from the width of one spectral line in 2004. Forty-one billion, from another line on the same object in 2019. The method itself is good to a factor of about four.

Two shadows. Both estimates. And Phoenix A, the hundred-billion one. That number is a sentence in a 2016 paper: the hole might have a mass of the order of a hundred billion Suns, judging by the energy the cluster gives off. Nobody weighed it. We draw it as a ghost, with no bar, because no bar is possible. Pull back again, from light-weeks to a hundred thousand light-years, and the Milky Way comes into view.

Its disc, measured, is eighty-seven thousand four hundred light-years across, give or take three thousand six hundred. The round hundred thousand is a convention. Now the biggest galaxy. IC 1101, about a billion light-years away. Galaxies have no edge. So a galaxy's size is where you stop counting. By the standard brightness cut, IC 1101 is about four hundred to five hundred and fifty thousand light-years across.

In 2026, an ultra-deep image traced its stars out to a radius of eight hundred and fifty thousand light-years. One point seven million across. Two edges, same galaxy: the bright one and the faint one. That is the error bar, made of starlight. The six million you may have read? No measurement behind it. Probably a 1991 halo radius mistaken for a diameter.

Porphyrion is bigger still: two jets of plasma from one black hole, twenty-three million light-years end to end. Plasma, not stars. And one thing every number in this act shares: these distances come from the expansion of the universe, and the two best measurements of that rate disagree by eight percent. Every size here stretches or shrinks with that argument. Now the one fast move. Out, past our galaxy, past Andromeda, until galaxies become dots. Forty-four thousand five hundred and ninety-nine of them, every one with a measured distance.

They are not spread evenly. They gather in filaments and walls. And between the walls are the holes. Three hundred million light-years out, we stop. The biggest hole. The Boötes void. Found in 1981: a region about two hundred and eighty million light-years across, where sixty galaxies sit where two thousand would be normal.

Not empty. Nearly empty. And the edge of a void is a choice. Draw the line where the density drops below a fifth of normal, and you get one size. On the same survey, three different void-finders found one thousand one hundred and sixty-three voids, five hundred and thirty-one, and five hundred and eighteen. Some say our own neighbourhood sits in a hole two billion light-years across, forty percent emptier than average. Supernova distances say it cannot be more than twenty-seven percent.

Disputed. Above the voids, the structures. Laniakea, our home supercluster, about five hundred and twenty million light-years, defined by where galaxies flow. The Sloan Great Wall, one point four billion light-years. And here is a line. Around one point two billion light-years, the universe is supposed to become smooth. Any structure bigger than that is not a record; it is a problem for the whole model.

Beyond the line, we draw ghosts. The Giant Arc, three point three billion light-years. The Hercules-Corona Borealis wall, up to ten billion. A 2025 paper ran the Arc's own algorithm on random dots, and found arcs. The wall rests on fourteen gamma-ray bursts. The people who found them and the people who doubt them are both still publishing. Which leaves the one biggest thing with a clean number.

The observable universe. Forty-six and a half billion light-years in every direction. Ninety-three billion across. About five hundred and fifty sextillion miles. A model and a definition, and we label it as one. Now the trap this whole episode has been avoiding. Volume.

Cut a radius in half, and you lose seven-eighths of the volume. That is how five billion Suns inside UY Scuti became three-quarters of a billion, without the star changing at all. So we say diameters, in miles, with the bar on. Volumes we leave to the people who are sure. Richard Feynman put it in one line: it is scientific only to say what is more likely and what less likely. The biggest star is probably about fourteen hundred times the Sun's radius. The heaviest black hole we have weighed, twenty to forty billion Suns.

The biggest galaxy depends on where you stop counting light. And the biggest structure is an argument. That is not a weaker answer. It is the measured one. Now fall all the way back down the ladder. Past the voids, the galaxies, the black holes, the giant stars, back home. We have measured stars wider than Jupiter's orbit, and black holes wider than the whole solar system.

But the most worrying rock astronomers have ever rated was about eleven hundred feet wide. Its name is Apophis. In 2004, for a few days, its odds of hitting Earth were two point seven percent. Next time: what it would have done, and why it misses. Next: What if Apophis hit your city?