A black hole that eats quickly ought to be a poor jet builder. Matter falling in at a high rate settles into a thin, cold disc, and standard theory says such discs struggle to hold the large-scale magnetic fields needed to fling material out at close to the speed of light. Yet some of the hungriest black holes known fire powerful jets anyway, and nobody has been able to say exactly where a black hole jet of that kind begins. A team in Japan now reports that, in one such galaxy, the disc simply stops short, and the jet is born in the hot gas that fills the gap.
The galaxy in question is 3C 120, an active galaxy in the constellation Taurus whose central black hole weighs roughly 63 million Suns. The study, led by Hirofumi Noda of Tohoku University and Kazuhiro Hada of Nagoya City University, was published in The Astrophysical Journal Letters on 19 October 2026. It pairs about 200 kiloseconds (roughly 55 hours) of data from XRISM, the X-ray Imaging and Spectroscopy Mission led by JAXA with NASA, with radio and millimetre images from three networks of telescopes. The X-ray observations were made in March 2025.
Everything hinges on a single spectral line. Iron atoms in the disc, lit by intense X-rays from just above it, glow at a characteristic energy. Close to a black hole that glow is smeared: gas orbiting at a large fraction of the speed of light is Doppler shifted, and light climbing out of the gravity well loses energy. The further inwards the disc reaches, the broader and more lopsided the line becomes. XRISM’s Resolve instrument, a microcalorimeter that measures each incoming X-ray by the tiny amount of heat it deposits, recorded that line in fine detail.
Its shape tells the story. According to the team, the cold disc ends at about 20 gravitational radii, a unit equal to half the event horizon radius of a non-spinning black hole. Inside that edge the gas appears to puff up into a thick flow heated to hundreds of millions of degrees, as Tohoku University’s release puts it. A model without this relativistically smeared component fits the X-ray spectrum badly. Adding it removes the misfit.
Radio telescopes supply the other half of the case. On 26 and 27 April 2025, 18 stations of the Global Millimeter VLBI Array observed 3C 120 at a wavelength of 3 mm, linking dishes thousands of kilometres apart to reach a resolution of around 50 millionths of an arcsecond. The Very Long Baseline Array and the East Asian VLBI Network traced the jet further out at longer wavelengths. Following the jet’s narrowing, parabolic shape back towards the black hole, the team finds it was launched from within about 10 gravitational radii. The jet fits inside the hole in the disc.
That geometry addresses a long-standing tension. The favoured explanation for jets, the Blandford-Znajek process, has a spinning black hole wind up magnetic field lines threaded through it, turning rotational energy into outflow. It sits comfortably with slow eaters such as M87, whose black hole shines at around a millionth of its Eddington limit, the luminosity at which the push of outgoing radiation balances the pull of gravity. 3C 120 shines at roughly 10 to 20 per cent of that limit. The new result suggests appetite may matter less than geometry: what counts is whether a thick, hot flow exists close in to hold magnetic field in place.
There are limits to what one galaxy can show. The inner edge comes from fitting the iron line with a phenomenological reflection model, a mathematical description rather than a full physical simulation, so its exact value depends on the assumptions built into that model. The authors present the Blandford-Znajek link as a possibility, not a demonstration, and the longer wavelength radio images were stacked from observations spread over several months. To test whether the pattern holds, the team has begun coordinated XRISM and radio campaigns on a more massive black hole. A preprint of the paper is freely available for anyone who wants the spectral fits in full.
If 3C 120 is typical, the hungriest black holes launch jets in much the same way as the quietest ones. Both need a hot, swollen flow near the horizon; the fast eaters simply keep it in a smaller space, hemmed in by a disc that reaches much closer to the horizon.
On this view the jet does not grow out of the disc at all. It grows out of the space where the disc gives up.
