The Repository Absolute Magnitude Space

S301: The Fastest Star in the Galaxy Has Two Orbits

September 16, 2026 · Runtime 12:07 · Watch on YouTube

S301 falls past the Milky Way's black hole at 25,000 km/s, eight per cent of the speed of light, every 8.7 years. No instrument has measured that speed. The discovery paper has nineteen positions, one brightness and no spectrum, and so it publishes two orbits.

Transcript

25,000 km/s, never measured 0:00

There is a star at the centre of our galaxy that falls past the black hole at twenty five thousand kilometres per second. Eight per cent of the speed of light. It comes around every eight point seven years, and at its closest it is nearer to the black hole than Saturn is to the Sun. It is the fastest star known in the Milky Way. And nobody has ever measured its speed.

Nineteen positions, no spectrum 0:25

What the instrument has returned for it is nineteen positions on the sky and one brightness. No spectrum, and so no velocity at all. This episode ends on the number the instrument actually returned for S three oh one, and on why the paper that announced it had to publish two orbits.

Last episode, a sixteen year orbit weighed the black hole at the centre of the Milky Way, to a quarter of a per cent. That orbit could not feel one thing about it: its spin. In August two thousand twenty six, the same collaboration published a star that can.

GRAVITY: four 8 m telescopes as one 1:03

The instrument is GRAVITY, at the Very Large Telescope in Chile. It combines the light of four eight metre telescopes into one, with fifteen times the sharpness of any one of them. An upgrade, GRAVITY plus, made it between ten and a hundred times more sensitive.

How a star on no list was found 1:22

In spring two thousand twenty three the team turned those combined fringes into an image, and found a point of light that was on no list, fifteen thousandths of a second of arc from the black hole. Once they knew where to look, they found it again in their own data from two thousand twenty one, and from two thousand seventeen.

Combined that way, the array resolves a few milliarcseconds, about the size of a car seen on the Moon. And according to the team, Paranal is the only observatory in the world with four eight metre telescopes that can work together as one interferometer.

Magnitude 19.3: two billion times fainter than Betelgeuse 1:58

It is faint. Magnitude nineteen point three in the infrared, two billion times fainter than Betelgeuse. That is part of why a star this close to the black hole went unseen for so long. It sits among much brighter neighbours, and the signal had to be pulled out of their glare.

Fifteen thousandths of a second of arc, at the distance of the galactic centre, is about a hundred and twenty astronomical units across the sky. Four times Neptune's distance from the Sun, seen from twenty seven thousand light years away.

In the end there are nineteen positions, spread over more than eight years. The closest pass came in early two thousand twenty three, just before anyone knew the star was there. The next one is due in two thousand thirty one.

What nineteen positions look like 2:49

This is what those positions look like. Each one is the star's offset from the black hole, measured on the sky, and together they trace an ellipse. That ellipse is the measurement. Everything else about S three oh one is built on top of it.

The chain: an angle, and S2's mass 3:07

Here is the chain. The size of the ellipse is an angle: eighty three milliarcseconds, plus or minus zero point seven. To turn an angle into a speed you need a mass and a distance, and this paper measured neither. It fixed them at the values from S two's orbit, four point two nine seven million suns at eight thousand two hundred and seventy seven parsecs. The number the last episode took apart.

The speed is required, not observed 3:35

With those fixed, Kepler does the rest. Eighty three milliarcseconds at that distance is six hundred and eighty seven astronomical units. The orbit is so stretched that the closest approach is under two per cent of that, about twelve astronomical units. And a star that close to four million suns has to be moving at twenty five thousand kilometres per second. The speed is not observed. It is required.

Even the period is the angle 4:04

Even the eight point seven years is tied to the angle. With the mass fixed, the period has no freedom of its own. Its error bar is one point two seven per cent, and the angle's is zero point eight four. One and a half times, which is Kepler's law, written as an uncertainty.

For scale: the Earth goes around the Sun at thirty kilometres per second. S two, at its closest in two thousand eighteen, reached seven thousand six hundred and fifty. S three oh one, on paper, reaches more than three times that.

A position has no depth 4:42

Now the catch. A position on the sky is two numbers: left and right, up and down. It has no depth. From positions alone you cannot tell whether the near side of the orbit is tilted toward us, or away from us.

Two orbits, equally valid 4:59

So the paper publishes two orbits. The same ellipse on the sky, mirrored in depth, their orientations about a hundred and eighty degrees apart. In the authors' words, the two are equally valid. The data cannot choose.

There is one slow way out. Light from the far side of an orbit reaches us later than light from the near side, and that delay could, in principle, tell the two orbits apart. With nineteen positions it does not, yet. The two fits come out indistinguishable.

No spectrum, no velocity 5:35

For S two, that question was settled more than twenty years ago, by a spectrograph. Light from a star moving toward us shifts blue, and away from us shifts red, and that shift is a velocity in kilometres per second, with a sign. For S three oh one there is none. The team searched deep spectra from the ERIS instrument, and could not detect the star at all.

Where each headline number comes from 6:25

A spectrum would carry more than a direction. At S two's closest pass in two thousand eighteen, relativity added about two hundred kilometres per second to its redshift, and that was measured. For S three oh one, the same simple estimate gives roughly two thousand. None of it was recorded in two thousand twenty three. There was no spectrum to record it in.

So put the headline numbers in one column, and ask where each one came from. The angle is measured, to under one per cent. The mass is borrowed from another star. The closest approach rides on one minus the eccentricity, known to about six per cent. And the famous speed is computed from all of them, to about three.

Even the kind of star is a conversion 6:49

Even what kind of star it is, is a conversion. Nineteen point three magnitudes, corrected for the distance and for the dust in front of the galactic centre, fits a star a little hotter than the Sun, between one point one and one point five solar masses. And the paper's best account of how it got there is a binary, torn apart by the black hole.

One star of the pair was captured onto this orbit. The other would have been flung out of the galactic centre altogether, as a hypervelocity star. None of that is in the nineteen positions. It is what a brightness becomes, once a distance, a dust correction and a model of stars are added.

Frame dragging: why the orientation matters 7:33

Here is why anyone cares about the orbit's orientation. A spinning black hole drags space around with it. An orbit that passes close enough is slowly twisted, its plane turning a little on every lap. It is called the Lense Thirring effect, and this is the first star found where it is strong.

2 degrees from mass, 0.11 from spin 7:53

The paper puts numbers on both twists. From the mass alone, the orbit's long axis turns by about two degrees every lap. The spin adds up to zero point one one degrees to that same turn, and that is the ceiling. It is multiplied by the spin, which runs from zero to one, and by the cosine of the angle between the spin axis and the orbit.

And there is the knot. That cosine needs the angle between the spin and the orbit, which means it needs the orbit's orientation in three dimensions. Which is exactly what nineteen positions cannot give. The same spin would show up differently in each of the two published orbits.

What a spin measurement needs 8:37

The paper's forecast assumes two things. Positions to a tenth of a milliarcsecond, and a velocity along the line of sight to one kilometre per second. With both, their simulated data pin the spin to better than plus or minus zero point two, within about a decade. The authors call that a fair chance.

The spectrum waits for the ELT 8:59

That velocity has to come from a spectrum of a nineteenth magnitude star, beside the brightest stars at the centre of the galaxy. The plan is MICADO, on the Extremely Large Telescope, whose thirty nine metre mirror is still being built on Cerro Armazones, in the Atacama.

The next closest pass is due in two thousand thirty one. That is when the star moves fastest, and when the relativistic effects in this episode are strongest. Whether a spectrum is ready to catch it depends on telescopes that are still being finished.

A disc can imitate the spin 9:35

And even then, the spin is not the only thing that can turn an orbit's plane. Any mass around the black hole that is not perfectly round, a flattened disc of stars or gas, pulls on the orbit too. A companion paper from the same collaboration puts numbers on how well that can imitate the spin.

For S two, in their disc model, the ordinary Newtonian twist is about fourteen times the spin signal, which is why S two could never feel it. For S three oh one it is about one and a half. Closer, but still larger than the thing being measured.

What S three oh one does have is proximity. It passes at two hundred and eighty gravitational radii, ten times closer than S two, and its spin signal per lap is about twenty nine times stronger. The plan is to subtract the disc using S two, S fifty five and S thirty eight, which feel the disc and barely feel the spin.

A speed nobody has clocked 10:36

So the fastest star known in the galaxy has a period, and an ellipse drawn to under a per cent. It has a speed that nobody has clocked, and an orientation that is still one of two. The one number that would settle it, a velocity in kilometres per second, is waiting on a telescope that is not finished.

S two was followed for three decades, with spectra, before its orbit weighed a black hole to a quarter of a per cent. S three oh one has eight years of positions. The difference is not the instrument. It is time, and the missing half of the measurement.

So the practical version. When a star's speed is quoted, ask whether anyone measured a line shift. If nobody did, the speed is an ellipse multiplied by somebody else's mass and distance, and the orbit may still come in two.

83.0 milliarcseconds. Next: Betelgeuse 11:30

Eighty three milliarcseconds, plus or minus zero point seven. The size of one orbit around the black hole at the centre of the Milky Way, drawn from nineteen positions by an instrument that has never seen the star's spectrum. The speed, the closest approach and the spin all hang from that one angle. Next time: the star this episode used as a ruler. Betelgeuse, and the winter it faded to magnitude one point six. Absolute Magnitude Space. Every episode ends in a measured number.

Description and sources

GRAVITY, on ESO's Very Large Telescope, spotted S301 in spring 2023, fifteen milliarcseconds from Sagittarius A*, and found it again in its own data from 2021 and 2017. What the instrument returned is nineteen positions on the sky and a brightness, K = 19.3. The orbit's size is an angle, 83.0 +/- 0.7 milliarcseconds; the fit fixes the black hole's mass and distance at the values from S2's orbit, and Kepler does the rest: 687 AU, a closest approach near 12 AU, and a required 25,000 km/s. The team searched deep ERIS spectra and could not detect the star, so there is no radial velocity and no depth: two mirror-image orbits fit equally well. That matters for the spin. The Lense-Thirring share of the orbit's turn is at most 0.11 degrees per lap, against about 2 from the mass alone, and reading it needs the orbit's 3-D orientation, a velocity to 1 km/s from the Extremely Large Telescope, and a way to subtract a flattened disc whose pull, in one model, is still 1.5 times the spin's. The episode ends on the number the instrument returned: 83.0 +/- 0.7 milliarcseconds.

PRINT-READY, FROM THIS CHANNEL

The Conversion Chain, Vol. 01 — the method, and four numbers taken apart with it

https://therepository.gumroad.com/l/the-conversion-chain

PRIMARY SOURCES

GRAVITY+ Collaboration (Abd El Dayem et al. 2026). Discovery of a star sensitive to the spin of Sgr A*. Nature, doi 10.1038/s41586-026-10894-w; arXiv:2607.12664 -- 19 positions, K = 19.3 +/- 0.3, a = 83.0 +/- 0.7 mas, P = 8.68 +/- 0.11 yr, e = 0.9832 / 0.9824 (two orientations), mass and R0 fixed at 4.297e6 suns, 8,277 pc; no radial velocity (ERIS); Schwarzschild 2.0/1.9 deg, Lense-Thirring 0.11 deg chi cos xi per revolution.
ESO eso2612 (19 Aug 2026) -- 25,000 km/s, 8.7 yr, 1.78 billion km; spotted spring 2023, traced back to 2017; next pericentre 2031; 15 times the resolution of one 8 m telescope.
Max Planck Society press release (19 Aug 2026) -- GRAVITY+ sensitivity gain 10 to 100; a few milliarcseconds, a car on the Moon; two billion times fainter than Betelgeuse.
Piran et al. (2026). S301 and friends: Measuring the spin of Sgr A*. arXiv:2607.24931 -- pericentre 280 r_g, e = 0.9825; LT precession 9.6e-4 vs 3.3e-5 rad per orbit (x29 vs S2); disc/LT ratio ~1.5 (S301), ~14 (S2); reference stars S2, S55, S38.
GRAVITY Collaboration (2022). A&A 657, L12. arXiv:2112.07478 -- M = 4.297 million suns, R0 = 8,277 pc; a(S2) = 124.95 mas.
GRAVITY Collaboration (2018). A&A 615, L15. arXiv:1807.09409 -- S2 pericentre May 2018: 120 AU, ~7,650 km/s, ~200 km/s relativistic redshift.

ARCHIVE PLATES — public-domain documents, with author, date, archive and licence.

Simulation of the orbits of stars around the black hole at the centre | ESO | cc-by-4.0
Yepun to the Universe | ESO | cc-by-4.0
Zooming into the black hole at the centre of our galaxy | ESO | cc-by-4.0
The most powerful laser guide star system in the world sees first ligh | ESO | cc-by-4.0
Unit Telescope 1 (UT1) in Action | ESO | cc-by-4.0
Effect of a spinning black hole on the orbit of the S301 star | ESO | cc-by-4.0
CRIRES mounted on Antu | ESO | cc-by-4.0
Paranal time-lapse | ESO | cc-by-4.0
Disruption of a binary star close to a black hole | ESO | cc-by-4.0

... and 6 more plates, each credited on screen in the video and listed in this project's manifest.

Absolute Magnitude takes one number the internet repeats and asks what an instrument actually returned. Every episode ends on a measured figure with its source on screen, and says plainly which part of it was converted afterwards.

No AI-generated imagery is used on this channel. Every frame of archive footage is real, publicly released material, credited on screen with its licence; every diagram is drawn in code from the numbers in the cited papers.

Corrections are pinned, not quietly edited.

#SagittariusA #BlackHole #S301 #Astronomy