Astronomers Pinpoint Source of Black Hole's 3,000-Light-Year Jet Stream

In 1918, astronomer Heber Curtis examined a photographic plate of the galaxy Messier 87 and noticed what he described as a "curious straight ray" apparently connected to its nucleus [1]. He had no way to know that this faint streak was a relativistic jet — a column of magnetized plasma hurtling outward at nearly the speed of light, stretching over 3,000 light-years into intergalactic space. More than a century later, an international collaboration has traced that jet back to its origin: a compact region just 0.09 light-years from the supermassive black hole at M87's center [2][3].

The findings, published January 28, 2026 in the journal Astronomy & Astrophysics, represent the first direct observational constraint on where M87's jet begins — a result made possible by adding two new telescope stations to the Event Horizon Telescope (EHT) network and analyzing data from its April 2021 observing campaign [4].

The Discovery: A Compact Source Near the Shadow

The research team, led by Saurabh, Hendrik Müller, and Sebastiano von Fellenberg of the Max Planck Institute for Radio Astronomy, confronted a puzzle that had emerged from earlier EHT data. When they modeled the bright ring of radio emission surrounding M87* — the black hole's "shadow," first imaged in 2019 — they found that the ring alone could not account for all the detected radio flux. Previous observations had noted a discrepancy of roughly 1.0 to 1.4 jansky between the total measured flux and what the compact ring structure could explain [4].

By fitting the 2021 EHT data with models that included both the ring and an additional Gaussian component, the team identified a compact emission source located approximately 320 microarcseconds east and 60 microarcseconds south of the ring center. This source, with a flux density up to about 60 millijansky and a full width at half maximum of roughly 180 microarcseconds, sits in the southwestern direction — aligned with M87's large-scale jet orientation [4]. Its position corresponds to a physical distance of approximately 5,500 astronomical units (about 0.02 parsecs) from the black hole, matching theoretical predictions for where the jet should be launched [4][5].

"This study represents an early step toward connecting theoretical ideas about jet launching with direct observations," said lead author Saurabh of the Max Planck Institute [3]. His colleague Sebastiano von Fellenberg added: "We aren't just calculating where these structures are anymore; we are moving toward being able to image them directly" [6].

What Made This Possible: An Upgraded Telescope Network

The 2021 observations benefited from a significantly expanded EHT array. Two stations were integrated into the network for the first time: the 12-meter Kitt Peak Telescope in Arizona and the NOEMA (Northern Extended Millimeter Array) observatory in the French Alps [4][7]. These additions were not about raw resolving power — the EHT's longest baselines already provided angular resolution of roughly 20 to 25 microarcseconds at 230 GHz. Instead, they filled critical gaps in the array's spatial frequency coverage (known as the u,v plane), providing sensitivity to structures at intermediate angular scales [4].

Specifically, the baseline between NOEMA and the IRAM 30-meter telescope at Pico Veleta in Spain probes structures at roughly 250 microarcseconds, while the baseline between the Kitt Peak 12-meter and the Submillimeter Telescope (SMT), also in Arizona, is sensitive to scales of about 2,500 microarcseconds [4]. These intermediate baselines bridge the gap between the EHT's finest resolution (which sees the ring) and the shortest baselines (which detect extended jet emission but cannot distinguish fine structure near the black hole).

The full 2021 array comprised eleven stations: ALMA and APEX in Chile, the IRAM 30-meter telescope in Spain, NOEMA in France, the SMT and Kitt Peak 12-meter in Arizona, the JCMT and SMA in Hawaii, the LMT in Mexico, the Greenland Telescope, and the South Pole Telescope [4][7]. The observing campaign ran April 9–19, 2021, with M87* observed on five nights. The analysis focused primarily on data from April 18, when conditions were optimal [4].

The EHT collaboration draws funding from multiple international sources. The U.S. National Science Foundation (NSF) has been a primary funder for over a decade, with additional support from the European Research Council (ERC) and funding agencies across East Asia [8]. The ALMA phasing system — critical for turning the 66-dish ALMA array into a single ultra-sensitive EHT station — was funded through the ALMA North America Development Fund and the NSF Major Research Instrumentation Program [8].

Source: Various VLBI publications

Data as of Jan 28, 2026CSV

A Century of Sharpening the View

The path from Curtis's photographic plate to the 2021 EHT result was neither straight nor fast. Understanding why requires grasping the technical barriers involved.

M87 lies approximately 53 to 55 million light-years from Earth [6][9]. At that distance, the black hole's event horizon subtends an angle of only about 10 microarcseconds — roughly the size of a coin on the Moon as seen from Earth. Resolving structures at the jet's launching region demands angular resolution measured in tens of microarcseconds, achievable only through very long baseline interferometry (VLBI) at millimeter wavelengths.

Early VLBI observations at centimeter wavelengths in the 1990s achieved resolution of roughly 1,000 microarcseconds — enough to trace M87's jet on parsec scales but far too coarse to see its base [1]. By the mid-2000s, 43 GHz observations with the Very Long Baseline Array (VLBA) pushed resolution to about 300 microarcseconds, revealing the jet's limb-brightened structure and a broad opening angle close to the nucleus [10]. A key advance came in 2016, when 86 GHz VLBI observations using the VLBA combined with the Green Bank Telescope resolved the jet base down to approximately 10 Schwarzschild radii, demonstrating that the jet's collimation began very close to the black hole [11].

The EHT's 2017 observations at 230 GHz — the data behind the iconic 2019 black hole image — achieved roughly 25 microarcseconds, enough to image the ring but not to separate it from the jet's launch region [12]. In 2018, GMVA+ALMA observations at 86 GHz confirmed a ring diameter of approximately 61 microarcseconds and simultaneously imaged the innermost jet, connecting the accretion structure to the outflow for the first time in a single image [13]. The 2021 enhanced EHT observations then provided the intermediate-scale sensitivity needed to isolate the jet base as a distinct component.

Each of these steps required not just better telescopes, but better electronics (wider recording bandwidths), better calibration techniques (to handle atmospheric turbulence at millimeter wavelengths), and better computational methods (to reconstruct images from sparse interferometric data). The timescale reflects the cumulative difficulty of all these advances happening in concert.

Two Theories of Jet Launching — and What This Result Says

Astrophysicists have proposed two primary mechanisms for how black holes launch jets, both rooted in the interaction between magnetic fields and rotating matter.

The Blandford-Znajek (BZ) process, proposed in 1977, extracts energy directly from the black hole's spin. Magnetic field lines, anchored in the surrounding accretion flow, thread the event horizon. As the black hole drags spacetime around it (a phenomenon called frame-dragging), these field lines twist, generating an outward electromagnetic flux — a Poynting jet — that carries energy away from the black hole [14][15]. The jet in this model originates from the ergosphere, the region just outside the event horizon where frame-dragging is strongest.

The Blandford-Payne (BP) process, proposed in 1982, instead accelerates material magneto-centrifugally from the surface of the accretion disk. Magnetic field lines anchored in the disk fling material outward as the disk rotates, like beads on a spinning wire [14]. In this model, the jet's footpoint is located in the disk, at some distance from the event horizon.

The distinction matters because the two processes predict different launch radii. A BZ jet should originate very close to the black hole — within a few gravitational radii. A BP jet should originate from the disk surface, which could be somewhat farther out. The compact source detected by the 2021 EHT observations sits at a projected distance of roughly 0.02 parsecs (about 5,500 AU) from the ring center [4]. However, the paper notes that the parabolic jet collimation profile observed in M87 — where the jet width scales with distance raised to the power of approximately 0.5 — is consistent with both mechanisms [4]. The current spatial resolution cannot definitively separate the two models because both predict structures that overlap at the scales probed.

Professor Bart Ripperda, commenting on the findings, noted that "the jets are theorized to be launched tapping the rotational energy of the black hole through electromagnetism" — a description that favors the BZ interpretation, though it does not exclude disk-driven contributions [6].

General relativistic magnetohydrodynamic (GRMHD) simulations suggest that in practice, both mechanisms may operate simultaneously, with a fast, spine-like BZ jet surrounded by a slower, disk-driven BP wind [14]. Distinguishing between them observationally will require multi-frequency VLBI imaging that can map the magnetic field geometry and velocity structure at the jet base with sub-gravitational-radius precision — a goal for the next-generation EHT.

The Black Hole: Mass, Spin, and Sensitivity to Assumptions

M87* has a well-constrained mass of approximately 6.5 billion solar masses, established through both stellar dynamics measurements and the EHT ring diameter [9][12]. Its spin is less certain but appears to be high. A 2023 study tracking the precession of M87's jet base over two decades of VLBA monitoring estimated a spin parameter of at least 0.8 (where 1.0 is the theoretical maximum), with some analyses suggesting values as high as 0.9 [16][17].

This matters for jet power calculations. In the Blandford-Znajek framework, the jet's electromagnetic luminosity scales roughly as the square of the black hole's spin parameter and the square of the magnetic flux threading the horizon [14]. A spin of 0.9 versus 0.5 would change the predicted BZ jet power by more than a factor of three. If M87*'s spin turned out to be significantly lower than current estimates — say, below 0.3 — the BZ mechanism alone would struggle to account for the observed jet kinetic power of approximately 10^44 erg/s (10^37 watts), and disk-driven processes would need to play a larger role [18][19].

Current estimates of M87*'s accretion rate range from 0.00004 to 0.4 solar masses per year, well below the Eddington limit — placing the black hole in a radiatively inefficient accretion state where jet production is expected to be efficient [16].

How Much Energy, and What Does It Do?

M87's jet carries an estimated kinetic power of roughly 10^44 erg per second — about 10 billion times the luminosity of the Sun [18]. This energy is deposited into the surrounding medium as the jet propagates, creating cavities and shock fronts in the hot X-ray-emitting gas that pervades M87's host cluster, the Virgo Cluster.

Source: Various astrophysical surveys

Data as of Jan 28, 2026CSV

This energy injection is central to what astrophysicists call AGN feedback. Without it, the hot gas surrounding massive elliptical galaxies like M87 would cool radiatively and collapse inward, fueling runaway star formation. Instead, the jet heats this gas, establishing a rough equilibrium between cooling and heating that regulates star formation over timescales of hundreds of millions of years [20][21]. X-ray observations of M87 show that the jet has carved out bubbles in the surrounding gas, and the mechanical power required to inflate these cavities — estimated at roughly 3 × 10^43 erg/s — provides an independent lower bound on the jet's energy output [18].

Whether this feedback is purely suppressive (preventing star formation) or sometimes constructive (triggering star formation through shock compression of gas clouds) depends on the geometry and power of the jet-medium interaction. Simulations and observations of other systems suggest both outcomes are possible, sometimes simultaneously in different regions of the same galaxy [21].

Is M87 Typical or Exceptional?

Roughly 10 percent of active galactic nuclei (AGN) are classified as radio-loud — meaning they produce powerful relativistic jets detectable at radio wavelengths [22]. Among the tens of thousands of known AGN, several hundred have jets that have been studied in detail through VLBI imaging. M87 occupies a somewhat special place in this population: it is one of the closest radio galaxies with a powerful jet, making it uniquely suited for high-resolution study.

In terms of jet power, M87 is moderate. Its estimated 10^44 erg/s places it well below objects like Cygnus A or the quasar 3C 273, which have jet powers of roughly 10^46 erg/s [18]. Its jet length of 3,000 light-years (about 1 kiloparsec) is also unremarkable — some AGN jets extend for millions of light-years, with one recently discovered pair spanning over 7 megaparsecs [23]. What makes M87 exceptional is its proximity and the mass of its black hole, which together yield an angular size large enough for the EHT to resolve structures near the event horizon.

This raises a legitimate question about generalizability. The jet-launching physics observed in M87 may or may not be representative of the broader AGN population. M87* is in a low-accretion, radiatively inefficient state, which differs from the high-accretion states seen in many quasars. Whether the same BZ/BP physics operates across this range of accretion rates is an open question that will require similar observations of other sources — Centaurus A, 3C 84, and eventually more distant AGN — to answer [22].

What This Observation Does Not Tell Us

A significant limitation of the current result concerns the jet's particle content. Astrophysical jets could be composed primarily of electron-positron pairs (a "pair plasma") or of electrons and protons (a "normal matter" plasma). The distinction has major implications for the jet's momentum, energy content, and interaction with the surrounding medium.

Analysis of M87's jet properties — combining VLBI surface brightness measurements, synchrotron emission modeling, and GRMHD simulations — has favored an electron-positron composition [24][25]. However, these conclusions rest on indirect arguments. The 2021 EHT observations, conducted at a single frequency (230 GHz), do not directly constrain the ion content of the jet. Probing particle composition would require polarization observations across multiple frequencies, detection of gamma-ray absorption features, or identification of spectral signatures unique to pair production — none of which this dataset provides [24].

Skeptics are right to note that pinpointing the jet's footpoint in one frequency band does not resolve this ambiguity. The compact source detected could be the base of a pair-dominated BZ jet, the root of a proton-loaded BP disk wind, or a blend of both. The spatial localization is a necessary but not sufficient step toward a complete physical model of jet launching.

The Research Pipeline: A Growing Field

Source: OpenAlex

Data as of Jan 1, 2026CSV

Academic interest in black hole jets has surged over the past decade. OpenAlex data shows that publications containing "black hole jet" have grown from roughly 1,659 papers in 2011 to a peak of over 6,100 in 2023, driven in part by the EHT's 2019 black hole image and subsequent Sagittarius A* results in 2022. The field has produced over 49,000 papers to date [26].

What Comes Next

The next-generation EHT (ngEHT) project aims to roughly double the number of stations in the array and add multi-frequency observing capability — allowing simultaneous imaging at 86, 230, and 345 GHz [7]. This would provide the spectral index and polarization maps needed to constrain the magnetic field geometry and particle content at the jet base. Future observations including ALMA, the LMT, and additional planned stations should substantially sharpen the picture of M87's jet launching region [5].

For now, the 2026 result closes one chapter of a century-long investigation while opening several new ones. The jet's approximate footpoint is located. The mechanism that powers it, the particles it carries, and whether these findings apply beyond a single convenient target — those questions remain.