Here on planet Earth, we normally think of our own home world as round, and yet it’s anything but a perfect sphere. If it were made of a perfect fluid, didn’t rotate, and were completely isolated from other masses that exert tidal forces on it (like the Sun and the Moon), it would be theoretically be massive enough to pull itself into a perfectly spherical shape, but that doesn’t line up with physical reality. The Earth instead is made of non-uniform material, has several interior layers that subduct and uplift, and is flattened into an oblate spheroidal shape: compressed at the poles and bulging at the equator. All told, the peak of Chimborazo in Ecuador is farthest from Earth’s center at 6384.4 kilometers, while the geographic North Pole in the Arctic Ocean is closest point on the atmospheric surface at just 6356.77 kilometers away.

All told, that means that Earth is “only” a sphere to a precision of 99.57%, with an 0.43% departure from perfect sphericity. Lower mass worlds, like moons, dwarf planets, asteroids and Kuiper belt objects, have even higher mountains and deeper craters on them relative to their world’s size, whereas the highest-mass planets in our Solar System, Jupiter and Saturn, rotate so quickly that their equatorial and polar radii differ by between 7-10%: much greater than Earth’s.

But the Universe is vast and varied, with objects much more massive — and hence, with much more gravity — than any planetary objects we know of. Which ones, among them all, are the closest to being a perfect sphere? The competition is presently between the slowest-rotating massive stars and the slowest-rotating neutron stars, with the ultimate winner having yet to be determined. Here’s what we know today.

First, before we look out into the deep Universe, there are two examples from the Solar System that are the “most spherical” of anything we’ve ever known of firsthand. The first is a natural example: the roundest natural object in the Solar System happens to be the Sun. It’s all three of:

• the most massive object in the Solar System (over 1000 times more massive than Jupiter),
• composed of a plasma, which acts as a perfect fluid covering its surface without mountains or depressions,
• and a very slow rotator, with its equator completing a single rotation once every 25 Earth days and its poles completing one every 33 Earth days.

All told, the Sun’s equatorial radius is only about 5 kilometers greater than its polar radius; compared to its average radius of around 696,000 kilometers, that means it’s a perfect sphere to 99.9993% precision, with only a 0.0007% deviation. (Venus is the roundest among the planets, but its high-peaked mountains deviate from perfect sphericity by 0.18%.)

In the laboratory, however, we’ve beaten the roundness of the Sun ever so slightly: by constructing the incredible niobium-coated quartz spheres that were built for the Gravity Probe B mission to precisely test Einstein’s General Relativity. These 1.5 inch (4 cm) diameter spheres had their surfaces polished to within 40 atoms (about 10 nanometers) of equidistance from the center point, for a perfect sphere of 99.99995% tolerance (0.00005% deviation from sphericity).

This quartz sphere, on the left, and a similar sphere after being coated in niobium, right, were exquisitely manufactured and polished to be the most perfect sphere ever made in a laboratory on Earth. Just 1.5 inches (4 cm) in diameter, every point on their surfaces is no more than 40 atoms width farther from the center than any other point: a record-setting tolerance of 0.00005% percent.

In fact, the Gravity Probe B spheres are so spherical that if you were to scale that sphere up to the size of Earth — with a mean diameter of 6371 km — the highest mountains and deepest trenches would have only about 3.2 meters (10 feet) in elevation separating them. If the Sun were as spherical as the Gravity Probe B spheres, the difference between polar and equatorial diameters would be just a third of a kilometer. If we want to do better than that, or better than the Sun, for that matter, we have to look for objects that are massive, slowly rotating, and behave as perfect fluids all at once. In space, there is no “laboratory polishing” that naturally occurs.

There are all sorts of candidates for what types of objects could be more spherical than our Sun, than the Gravity Probe B spheres, or than anything else we can find in all of the cosmos. Objects we’d want to consider include:

• small stars,
• mid-size (Sun-like) stars,
• giant stars,
• white dwarfs,
• neutron stars,
• and black holes,

as any of them could, with the right intrinsic properties, be potentially “the most spherical” of all. But there are two conflicting ideas at play: the tension between being more compact and having a higher surface gravity, and of conserving angular momentum and spinning faster the more closely distributed your mass is to your central axis of rotation.

Let’s explain.

Saturn, as photographed here by Cassini during the 2008 equinox, isn’t quite round (as it’s more of an oblate spheroid), but is in hydrostatic equilibrium. With its low density and rapid rotation, Saturn is the most flattened planet in the Solar System, with an equatorial diameter that’s more than 10% larger than its polar diameter. Its colors and “bands” are largely due to different atmospheric layers appearing dominant in visible light at different latitudes.

For any massive collection of matter that you can imagine, it’s going to take up a specific amount of volume. In the idealized case, that object wouldn’t rotate at all, but would have zero angular momentum and be a perfect non-rotator. (In our Solar System, Venus is closest to that; if it were made up of a perfect fluid, rather than rock with peaks, valleys, crags, volcanoes, and craters, it would be the roundest object of all.) If we idealize the situation, then it wouldn’t matter how large an object was or how much volume it occupied; the fact that it was a perfect fluid that only was shaped by gravity would be all that mattered, and if it didn’t rotate and was completely isolated, it would indeed be a perfect sphere: with every point on its surface equidistant from its center.

However, if you imagine any imperfection on it — a departure from perfect sphericity — that departure is going to be more exaggerated dependent on how strong or weak the surface gravity is at that point. The amount of height that a mountain gets, or the amount that the equator bulges (or the poles get compressed) for a rotating object, is inversely proportional to the acceleration of gravity at the object’s surface. Therefore, whether an object is solid or fluid-like, or whether it rotates slowly or quickly, you’d expect that the more compact it is (i.e., how small its radius is), the more spherical it tends to be.

This simulation of a red supergiant’s surface, sped up to display an entire year of evolution in just a few seconds, shows how a “normal” red supergiant evolves during a relatively quiet period with no perceptible changes to its interior processes. There are large departures on its surface from perfect sphericity at any given time: most dominantly due to their incredibly low surface gravity.

Credit: Bernd Freytag, Susanne Höfner & Sofie Liljegren

If you consider fluid-like objects (to avoid the “Venus problem” where rocky objects, even if they’re extremely slow rotators, will still have large solid departures from sphericity), their surface gravities can be estimated as follows.

• Jupiter-like planet: 25 m/s², or 2.5 Earth gs,
• Brown dwarf (failed star): 1000 m/s², or 100 Earth gs,
• Red dwarf (low-mass star): 2500 m/s², or 250 Earth gs,
• Sun-like star (intermediate mass star): 274 m/s², or 28 Earth gs,
• Red giant star (evolved Sun-like star): 5 m/s², or 0.5 Earth gs,
• Blue supergiant (high-mass star): 3 m/s², or 0.3 Earth gs,
• Red supergiant (largest stars of all): 0.004 m/s², or 0.0004 Earth gs,
• White dwarf (stellar remnant): 1,000,000 m/s², or 100,000 Earth gs,
• Neutron star (stellar remnant): 10¹² (one trillion) m/s², or 100 billion Earth gs,
• Black hole (stellar remnant): Not well defined, because black holes only have event horizons.

For “normal” stars, it’s clear that the smaller ones tend to have greater surface gravities, while the high mass ones and the giant ones have lower surface gravities.

But if you go to the more compact objects — white dwarfs, neutron stars, and black holes — you start to get much greater surface gravities. White dwarfs are approximately equal in mass to the Sun, but only equal in size to the Earth: compressing an incredibly large amount of mass to impressively large densities. Neutron stars are just as massive (or even slightly more massive), but might be only 10-20 km in radius or so: the size of a large city. Their surface gravities can be hundreds of billions of times as strong as Earth’s gravity. Black holes don’t have surfaces; they only have event horizons. The lowest-mass black holes would have the highest gravity at their event horizons, but it’s not really fair to include them because they aren’t “objects” in the conventional sense of having a surface to be rounded.

When a figure skater like Yuko Kavaguti (pictured here from 2010’s Cup of Russia) spins with her limbs far from her body, her rotational speed (as measured by angular velocity, or the number of revolutions-per-minute) is lower than when she pulls her mass close to her axis of rotation. The conservation of angular momentum ensures that as she pulls her mass closer to the central axis of rotation, her angular velocity speeds up to compensate.

However, there’s a disadvantage to these more compact, denser objects as well: their departures from sphericity aren’t only dependent on their surface gravity (inversely proportional), but are also dependent on how fast an object is rotating: on its rotational velocity, or the number of rotations it makes in a given amount of time. Just as a figure skater who starts spinning pulls their arms and legs in and begins spinning faster, a star (or stellar remnant) starts spinning faster and faster the more severely its mass gets concentrated close to its axis of rotation.

If we would take an object like our Sun, which rotates (let’s simply it and pretend that it’s uniform) at the rate of one complete rotation every 25 days, if we shrunk the Sun down to the size of the Earth, so that it were similar to a white dwarf, but kept its angular momentum (which is conserved) the same, it would complete a full rotation in only three minutes! If you went a step further and compressed it down to the size of a neutron star, just 0.3% the radius of Earth, you would find that it makes a complete rotation in just 2 milliseconds: similar to the period of the fastest rotating objects known, millisecond pulsars.

The extreme temperature differences between a world in full sunlight at its closest approach to its parent star versus nighttime and/or its farthest point in orbit on that same world can lead to extreme temperature differences, with extreme consequences for phase changes in matter. The faster a star rotates, the greater it bulges at its equator and the more it compresses at its poles.

Credit: T. Mikal-Evans (MPIA) / T. Müller (MPIA/HdA)

The amount that an object departs from perfectly spherical isn’t just inversely proportional to their surface gravity, but it’s also proportional to their angular/rotational velocity squared, meaning that if an object has a significant amount of rotational velocity to it, that can be more important than its compactness for determining its departure from sphericity. Neutron stars, while the most compact objects in the Universe with the highest surface gravities, can also be the fastest rotators, with their surfaces often rotating at 10% or more of the speed of light; the record-holder spins with its surface moving at an estimated 24% the speed of light!

However, not every star, white dwarf, or neutron star rotates quickly at all. For that reason, if we want to find the most spherical objects in the Universe, that’s where we should look: to the slowest rotator, where either the slowest rotating stellar object of all or the slowest rotating stellar remnant of all will likely be the closest to spherical of any object in the Universe. The figures to beat, remember, are 99.9993% spherical (0.0007% aspherical) for the Sun, and 99.99995% spherical (0.00005% aspherical) for the laboratory-created spheres designed for Gravity Probe B.

An accurate size/color comparison of a white dwarf (left), Earth reflecting our Sun’s light (middle), and a black dwarf (right). When white dwarfs finally radiate the last of their energy away, they will all eventually become black dwarfs. The degeneracy pressure between the electrons within the white/black dwarf, however, will always be great enough, so long as it doesn’t accrue too much mass, to prevent it from collapsing further. A slow-rotating white dwarf is much more spherical than a similarly slow-rotating normal star can be.

Credit: BBC / GCSE (L) / SunflowerCosmos (R)

For the slowest-rotating star known, we can point to the star KIC 11145123: a star only slightly larger than the Sun, at 1.57 solar radii, but that rotates extremely slowly: with a speed of just under 1 km/s, meaning that it would take approximately 100 days to complete a single rotation. With the parameters given, it would be approximately a factor of 10 more spherical than the Sun: about 99.99993% spherical, or just 0.00007% aspherical. The actual velocity could potentially be even lower, which means that this one star could even rival the spheres made for Gravity Probe B.

But there are more compact objects out there, including a white dwarf that rotates even more slowly: white dwarf WD 1532+129, which was just determined earlier here in 2026 to have a rotational period of 289 days, making it by far the longest-period stellar remnant known to humanity, with an even longer rotational period than Venus. Based on the white dwarf’s mass of 0.57 solar masses, it should have a radius of 6237 kilometers, or 0.896% the radius of the Sun. Its surface gravity is therefore approximately 77,000 Earth gs, and that equates to a flattening of about 0.6 parts-per-trillion: far, far more spherical than any star or even the Gravity Probe B spheres, by a factor of approximately one million.

With a very strong surrounding magnetic field, pulsars accelerate matter around them and collimate them in jets that get emitted around two poles. As the neutron star rotates, the emitted jets rotate, and each time one of the jets crosses your line-of-sight, you observe a pulse of emissions: typically at radio frequencies, but often also at other frequencies as well.

If we go to even more compact objects, we come to the slowest rotating neutron stars known. On the one hand, from direct radio measurements, we have the slowest rotating pulsar, PSR J0901-4046, which is unusual in many ways: it’s very young at just 5.3 million years, as derived from its substantial spin-change rate, but takes a whopping 75.885 seconds to make one rotation. This object is both more massive and far denser than white dwarf WD 1532+129, but rotates far more quickly. When you work out the math, it turns out that this neutron star is far more spherical than any star, but has an asphericity that’s about 100 times greater than the roundest white dwarf.

However, there is an even slower-rotating neutron star that isn’t a pulsar, but that is instead a suspected magnetar: ASKAP J1935+2148. It’s located quite far away in our own galaxy, at 15,800 light-years, but has a rotation period of nearly an hour: 3225.313 seconds, to be precise. Its pulses were only first detected in 2022, so it’s an extremely recent discovery, but when you work out the math for its suspected mass, size, and rate of rotation, you find that it has the smallest flattening of any known object: of just 44 parts per quadrillion. It’s approximately a factor of 13 more spherical than even WD 1532+129, making it the most spherical object within the known Universe at present.

This computer simulation of a neutron star shows charged particles being whipped around by a neutron star’s extraordinarily strong electric and magnetic fields. Most neutron stars rotate quickly, giving rise to extremely strong magnetic fields and remarkably strong signals. Although they’re far less radio-loud, more slowly spinning neutron stars should also exist, with the record holder for the slowest, ASKAP J1935-2148, taking a remarkable 53.8 minutes to complete one rotation.

Credit: NASA’s Goddard Space Flight Center

It’s remarkable, when you think about it: at the time that the Gravity Probe B spheres were constructed, they were the most perfect spheres in the known Universe. Neutron stars were theoretically more spherical, but nobody knew if slow-enough rotating neutron stars existed to be certain.

Now, here in 2026, we know of regular, everyday stars that are just as spherical as those Gravity Probe B spheres are, and significantly more spherical than our slowly-rotating Sun. But white dwarfs can be even more spherical: approximately a million times more spherical, meaning that their difference between polar and equatorial radii is about the size of one wavelength of blue light. At an even greater extreme, the slowest-rotating neutron star — ASKAP J1935+2148 — has an estimated departure from perfect sphericity by about the size of four hydrogen atoms: ten times smaller than the imperfections of the 1.5-inch (4 cm) diameter Gravity Probe B spheres, but for an object that’s approximately 20 kilometers in physical size.

So currently, the roundest, most spherical object in the known Universe is the slowest-rotating neutron star. But keep in mind: there are only a few thousand known neutron stars at present, out of an estimated 500 million of them in our galaxy alone. There are almost certainly more perfect spheres out there. It’s up to us to find them!

This article What natural object is the roundest in the Universe? is featured on Big Think.

Source link