The instant the surface gave way
When a star collapses, the first light of the explosion is released at the moment the shock wave reaches the surface and breaks out, and it lasts anywhere from seconds to hours. On 21 March 2026 the Einstein Probe satellite detected a soft X-ray flash designated EP260321a; two peer-reviewed papers announced by NSF NOIRLab identify that flash as the shock breakout of the supernova SN 2026gzf. The thermal X-ray peak is 1.0 × 10⁴⁵ erg/s at a temperature of 130 electronvolts. In the past two decades only one other such breakout has been caught in X-rays with confidence.[1]
The follow-up that came afterwards describes an event more violent than an ordinary core collapse. Brendan O'Connor and colleagues place the source at redshift 0.0344, a luminosity distance of 158 megaparsecs, and give ejecta of 2.0 solar masses with 0.45 solar masses of nickel. The velocity read from the width of the iron lines is 35,000 km/s at 3.6 days and 20,000 km/s by day 11. These are the numbers that suit a broad-lined Type Ic explosion of a star that had already shed its hydrogen and helium shells.[1]
The jet that was not there
Explosions in this class are usually spoken of together with a gamma-ray burst and a relativistic jet. Here neither appears. Jillian Rastinejad and colleagues report that radio follow-up carried out between 5.8 and 54.5 days rules out an on-axis jet, and that deep X-ray observations exclude a typical gamma-ray burst afterglow. The teams read this as a weak, mildly relativistic outflow choked inside the material of the star.[1]
The prosaic account deserves to be stated first. A jet may exist and point away from the line of sight; in that geometry radio follow-up would also stay silent. What separates the choked scenario is that an outflow trapped inside the star both delays and weakens the shock breakout, and that this fits the X-ray luminosity actually measured. The difference between the two accounts can be narrowed by watching the late radio emission of the same explosion: a jet pointing off axis reveals itself months later as it spreads.[1]
The count itself sets a limit
The second result Rastinejad and colleagues report reaches beyond one star. The Einstein Probe detection rate so far does not match the shock breakout frequency expected from Type Ic-BL supernovae. That means either these explosions emit less X-ray light than predicted, or some fraction of the fast transients the satellite sees comes from something else. A second and a third example separate the two possibilities.[1]