A massive star vanished. No supernova. No light show

A massive star in the galaxy next door spent ten million years burning. Then in 2014, it flickered. By 2023, it was just — gone.

No supernova. No fireworks. No light show that anyone on Earth or in orbit ever saw. Just gone. And in the spot where it used to be, astronomers think there's now a brand-new black hole — born without the death scream every other massive star delivers.

This is the story of M31-2014-DS1. The star that forgot how to die properly.

The full video, 11 minutes — https://youtu.be/hNuMVP-GhDQ

A massive star — anything more than about eight times the mass of the Sun — burns through its fuel in roughly ten million years. Then the core runs out of fuel, gravity wins, and the core collapses. The collapse fires a shockwave outward. The shockwave blasts the outer layers off into space. What's left behind is either a neutron star or a black hole, and the blast itself — the supernova — is bright enough, for a few weeks, to outshine the entire galaxy it sits in.

This is textbook physics. Baade and Zwicky proposed the mechanism in 1934. Every astronomy class teaches it. Massive stars explode. That is the rule.

The catch is in the energy budget. If the shockwave doesn't push hard enough on the outer envelope, the envelope just falls back in. No light show. The star quietly disappears, and a black hole is left behind in its place. The technical name is direct collapse, or a failed supernova.

And the math is more than theoretical. Astronomers have a problem with the count. When you tally the massive stars that should be exploding against the supernovae we actually see, the numbers don't match. Somewhere between sixteen and thirty percent of massive stars seem to vanish without a recorded explosion. This is called the red-supergiant problem, and the leading explanation is that the missing stars aren't vanishing — they're going dark quietly.

That hypothesis has been a statistical hole for decades. Nobody had a clean individual case. Until now, maybe.

In February 2026, a paper landed in Science. Lead author Kishalay De, at Columbia. Title: Disappearance of a massive star in the Andromeda Galaxy due to formation of a black hole. It's the strongest observational case for a quiet collapse to a black hole that anyone has ever published.

The star is M31-2014-DS1 — a hydrogen-depleted supergiant, about twenty times the mass of the Sun, sitting in the Andromeda galaxy two and a half million light-years out. The timeline runs like this. In 2014, the star brightened in mid-infrared light — a moderate outburst, not a supernova. Around 2016 it started to fade. By 2022 and 2023 it was essentially gone in visible and near-infrared: about one ten-thousandth of its original optical brightness. A faint mid-infrared glow remains, consistent with leftover heated dust around whatever's left.

No supernova was ever recorded at that location. Not a flicker. Not an alert.

The only reason we know about any of it is that an infrared telescope was patiently watching the sky on its own time. That telescope is NASA's NEOWISE — the infrared sky-survey extension of WISE, retired in 2024 after eleven years on orbit. De's team went back through the public NEOWISE archive and pulled the lightcurve. Then they layered on follow-up from JWST, Hubble, Keck, the Chandra X-ray Observatory, Palomar's Zwicky Transient Facility, and the Large Binocular Telescope. A multi-decade, multi-instrument stack converging on one pixel in Andromeda.

The picture that came out the other end: a star that brightened, faded, and vanished. With no supernova. And — critically — no luminous X-ray accretion disk where it used to sit. Chandra looked. There's nothing bright eating material in that spot. Exactly what you'd expect if a black hole had quietly formed and the leftover dust had mostly fallen back in without lighting up.

The field is not settled. Within weeks of the Science paper, two more papers dropped on arXiv.

The first — Beasor and colleagues, MNRAS-accepted — backed up the failed-supernova interpretation with fresh JWST observations and a persistent dust-enshrouded source at the same coordinates.

The second is the one worth pausing on. The title — I am not making this up — was The failed failed-supernova scenario of M31-2014-DS1. That paper argues the 2014 brightening was actually a luminous red nova — the flare you get when two stars in a binary spiral together and merge — and the "disappearance" in this reading isn't a star turning into a black hole. It's the merged remnant hiding behind a shell of dust the merger threw off.

Both papers are from 2026. Both teams are good. Both interpretations fit a chunk of the data. The community has not picked a winner.

If De is right, humans now have the cleanest observational record of a massive star turning into a black hole without exploding. The textbook case for the next generation. Strong evidence that a meaningful fraction of massive stars die this way. If the merger paper is right, the failed-supernova phenomenon is still real in theory — but the cleanest example just got demoted, and the search continues.

This finding took ten years to surface — the brightening was 2014, the disappearance was complete by 2023, the paper landed in 2026. The reason it took that long is that nobody knew to look. The data was sitting in an archive, waiting for somebody to ask the right question of it.

That changes from here. The Vera Rubin Observatory in Chile had its first light in 2025. Its survey, the Legacy Survey of Space and Time, scans the entire visible southern sky every few nights, built specifically to catch transients in close to real time. The Roman Space Telescope, planned for 2027, adds an infrared layer. The next failed supernova won't take a decade to confirm. It will be a press release within weeks.

Which means — if the failed-supernova interpretation of M31-2014-DS1 holds up — this is the last one of these we'll ever have to find in the rearview mirror.

We don't know yet whether M31-2014-DS1 is the cleanest case of direct collapse on record, or a stellar merger in disguise. Two papers, same year, same star, opposite conclusions. The answer comes from more observations, more time, and probably a couple more candidates.

What's not contested is the underlying claim. The red-supergiant problem is real. Massive stars are going missing from the supernova count, and the math says they're going somewhere. The leading explanation: somewhere between one in six and one in three of them collapse into black holes without ever announcing it.

Somewhere out there in the dark, the universe is making black holes without telling anyone. We are just learning how to listen.

Full 11-minute video — https://youtu.be/hNuMVP-GhDQ

Full source list — including NASA/JPL, Caltech, Columbia, Keck, AAS Nova, and the ApJ Letters follow-up — is pinned in the YouTube comment.

Protocol Thesis decodes one peer-reviewed paper at a time. Forwarded by a friend? Subscribe here.

Narration in the video is AI-generated (ElevenLabs). Research, script, and editorial judgments are human-authored.