OrbitalHub

A microchip sat on the stage of beamline 26-ID at Argonne National Laboratory on a Tuesday in early October. Instead of guiding a motorized stage through a hand-scripted raster, a researcher typed “find the interface between these two regions” into a chat box. A software agent read the instruction, parsed a freshly reconstructed X-ray image, picked a new point on the sample, asked the beamline to move there, looked again, and decided on a third step. After a handful of those scan-reason-move cycles, the agent locked onto the boundary the user wanted and switched itself into a higher-resolution scan mode. No operator wrote Python. The human in the room just talked to the microscope.

That experiment, run at the joint Center for Nanoscale Materials and Advanced Photon Source hard X-ray nanoprobe in the first week of October 2026, is the first time a working agentic AI has steered a synchrotron beamline end-to-end through natural language. The team behind it, drawn from Argonne, Lawrence Berkeley, Brookhaven, SLAC, and Oak Ridge national laboratories, is calling the broader effort SYNAPS-I, short for Synergistic Neutron and Photon Science–Intelligence, and plugging it into the Department of Energy’s Genesis Mission as a flagship demonstration of what an AI-augmented user facility can look like in 2026 (Argonne, October 5, 2026).

For most of the synchrotron’s seventy-year history, the bottleneck at a beamline has not been the X-ray source. It has been the human at the controls. A modern storage ring like the APS sends electrons whirling at seven gigaelectronvolts through a 1.1-kilometer ring of bending magnets, where they shed X-rays bright enough to resolve features a few nanometers across. The catch is that the data deluge starts the moment the shutter opens. At 26-ID, a single experiment at a third-generation nanoprobe writes millions of ptychographic diffraction patterns in a few hours, and every pattern needs a heavy iterative computation to turn back into a usable image. Without live reconstruction, the team on shift has to choose what to scan on the basis of intuition and a coarse preview, then fly home and discover, days later, that the region of interest was three millimeters to the left the whole time.

The SYNAPS-I project, launched in 2025 and led out of Lawrence Berkeley, set out to change that ratio. The first phase, publicly reported in April 2026, replaced the slow offline reconstruction with a real-time engine. The engine, a fine-tuned 2D vision transformer the team calls PtychoFM, ingests raw diffraction patterns as they stream off the detector and returns a reconstructed phase-and-magnitude image in seconds (Argonne, April 2026). The math itself is well understood: a coherent X-ray beam is scanned across the sample in overlapping circles, and the resulting diffraction patterns are stitched together in Fourier space using a difference-map or gradient-descent solver. What PtychoFM adds is a learned prior, a model trained on enough examples that it can skip most of the iteration and guess a plausible reconstruction from a single forward pass, then refine it on the fly as more data arrives. For a beamline scientist, that turns hours of post-experiment waiting into a real-time video feed of structure under the beam.

The second phase, announced in early October, is what made the headline demo possible. The team wrapped PtychoFM in an agentic layer that uses a multimodal large language model as its decision-maker, alongside Meta’s Segment Anything Model 3 (SAM3) as its image-segmentation eye. The agent reads each new image, calls SAM3 to label the structures it sees, and reasons about what to do next in plain English — or rather, in plain tokens. “Agentic AI solves this by interpreting the images reconstructed in real time by PtychoFM, our ptychographic model, leveraging the vision capability of a multimodal large language model,” said Ming Du, an Argonne computational scientist on the SYNAPS-I team, in the October release. The architectural pattern is now familiar from the wider AI world: an LLM plays the role of a slow, generalist planner, a smaller specialist model handles perception, and a set of tool calls translate decisions into beamline commands. Here the tools are the motor controllers of a nanoprobe stage and a reconstruction pipeline running on the Argonne Leadership Computing Facility.

The integration was not trivial. SAM3 is a foundation model of non-trivial size, and the team needed it available to every beamline in the DOE complex rather than baked into one workstation. The Argonne Leadership Computing Facility stood up a cloud-style inference service, the ALCF Inference Service launched earlier in 2026, and the SYNAPS-I team plumbed SAM3 into it. From the agent’s point of view, segmentation is a function call. From the operator’s point of view, asking the beamline “what’s at this spot?” and getting a labeled map back is now as cheap as a chat reply.

The microelectronics demo that closed the October announcement was deliberately mundane. The interface between two doped regions in a chip is a classic defect-investigation target; in 2026 it is also a target that is genuinely hard to localize across devices, because process variation shifts the boundary unpredictably from wafer to wafer. Aileen Luo, the Argonne assistant computational scientist who led the ptychography-model work, framed the result this way: real-time reconstruction “changes the way experiments are performed, enabling researchers to adjust measurements on the fly instead of discovering key features they missed after the experiment has ended” (Argonne, October 2026). The agentic layer extends that loop by closing it. The microscope now decides what to scan next, not just what it has already seen.

What changes because of this is not the physics. The X-ray wavelength, the optics, the sample stage, and the math of ptychographic reconstruction are all the same as a decade ago. What changes is the operator. A graduate student who used to spend a year learning EPICS control scripts can now describe what they want in a sentence and spend the saved time on the science. That is the same labor-shifting move that swept structural biology when cryo-EM got its own AI denoisers, and astronomy when adaptive-optics pipelines learned to schedule themselves. Beamlines are next. By 2030 it is reasonable to expect that a substantial fraction of routine scanning at a synchrotron user facility will be conversational: the operator sets a goal, the agent handles the geometry, the human reviews the result and asks for the next scan.

The cleverness here is mostly plumbing. PtychoFM does the hard imaging work. SAM3 does the segmentation. A multimodal LLM does the planning. The novelty of SYNAPS-I is that Argonne, Berkeley, Brookhaven, SLAC, and Oak Ridge wired those pieces into the same end-to-end loop on a working beamline, and the Department of Energy underwrote it through Genesis. That is harder to copy than any single model. It is also more useful.

On 22 September 2026, IonQ researchers posted a paper that quietly removes one of the most stubborn bottlenecks in quantum computing. Nicolas Delfosse, Min Ye, and Andrii Maksymov showed that a real-time decoder for fault-tolerant quantum error correction can run end-to-end on a single commercial CPU — specifically a 12-core Apple M4 Max laptop chip (arXiv:2608.25027). The benchmark held 408 logical qubits alive across more than a million T gates while adding only 0.02 percent of “stretch” time on top of the quantum run. Until now, every credible demonstration of a useful fault-tolerant quantum computer assumed a fleet of FPGAs, custom silicon, or beefy GPU clusters babysitting the qubits in real time. IonQ just argued, with measurements, that one is enough.

That claim matters more than it sounds. Quantum error correction has been the field’s longest-running promise and its largest engineering gap. A physical qubit loses coherence in microseconds; without constant repair, no useful computation finishes. The standard fix — a surface code — spreads each logical qubit across hundreds or thousands of physical qubits, measures parity checks (“syndromes”) between them, and feeds the syndrome stream to a classical decoder that has to guess which physical qubit flipped and push a correction back before the next gate fires. The classical side has to keep up. If decoding lags behind the quantum cycle, the quantum computer idles. If decoding fails, the logical qubit decays. This real-time constraint has shaped every architecture choice since Peter Shor introduced the first quantum code in 1995 and John Preskill coined the term fault tolerance four years later.

What IonQ did, in plain engineering terms, was throw out the assumption that fault tolerance requires exotic hardware. The team benchmarked a decoder on a workload they call MegaQuOp: roughly a million quantum operations, the threshold at which any algorithm starts to do something useful. Their dual-decoder architecture runs simultaneously on the memory blocks where logical data lives and on magic-state factories where the resource states that produce T gates are distilled. On a trapped-ion machine with a 1–5 millisecond cycle time — slower than superconducting systems but far cleaner per operation — the decoder finished its work in tens of microseconds per round, leaving the qubits essentially idle for the rest of the cycle.

The benchmarks themselves were a tour of the workloads any fault-tolerant computer will face. One was a measurement-induced phase transition circuit: 40 layers of random two-qubit gates compiled into Clifford and T gates at an average cost of 538 T gates per random gate, producing more than a million T gates across 102 logical qubits (arXiv:2608.25027, Tables I–IV). Another was a single sixth-order Trotter step of a disordered Heisenberg model: 64 sites on 102 logical qubits, then 266 sites on all 408. These are not synthetic stress tests; they map onto the early problems a useful quantum computer will tackle — chemistry, materials, condensed-matter simulation. Running them in real time, on commodity silicon, is the kind of result that moves a roadmap from “lab demo” to “engineering project.”

The decoder IonQ used is a variation on the minimum-weight perfect-matching (MWPM) algorithm that has dominated surface codes since Daniel Gottesman published the foundational results in 1997. Each syndrome measurement is a small piece of evidence about where an error happened; the decoder has to connect those pieces into consistent error chains and infer the most likely physical failure. The classical challenge is that the matching problem scales superlinearly with the number of qubits and gates, and the latencies required are measured in microseconds, not milliseconds. The way IonQ kept it tractable was to use a fast, sparse-graph formulation and to split the workload between two decoders running in parallel — one for memory, one for factory — so neither saturates the CPU. The result is a workload that fits comfortably on a 12-core laptop, with cores to spare.

This is also why the “single CPU” claim is not a marketing trick. The alternative — running the decoder on a rack of FPGAs, or on a dedicated ASIC, or on a GPU cluster — has been the design choice at IBM, Google, and Quantinuum. Each approach works in a lab. Each one also adds cost, power, and integration risk at exactly the moment the industry is trying to scale from hundreds to thousands of logical qubits. IonQ is betting that trapped-ion hardware, with its longer cycles and cleaner qubits, lets a conventional CPU do the work that other architectures have had to throw specialized hardware at. If the bet holds, the unit economics of a fault-tolerant machine shift sharply: the classical co-processor stops being a separate procurement line and starts being a server in the rack.

The result lands inside IonQ’s broader architectural plan, the Walking Cat scheme for fault-tolerant trapped-ion computing. Walking Cat is the company’s response to one of the hardest problems in the field: ion traps do not scale the way superconducting lattices do, because all the ions in a trap see the same control fields. Walking Cat threads the qubits through a chain of zones — memory, factory, measurement — moving them physically between regions while the controller keeps the surface code coherent. A lightweight decoder that fits on a CPU is a precondition for that scheme, because every zone transition generates syndromes that have to be processed before the next move. The 408-qubit benchmark is, in effect, the first measurement that says this scheme is implementable at scale.

What changes because of this. The next twelve months in fault-tolerant quantum computing will look different than the last twelve. Demonstrations that used to require a roomful of custom electronics will fit on a workstation; university groups will be able to reproduce them without privileged access to ASIC foundries. For IonQ, the result is a public milestone on the path from 256 physical qubits to “industrial-scale platforms controlling thousands of qubits,” to use the wording of the company’s 22 September announcement. For the field, it is one of those rare engineering results that removes a future constraint rather than adding one. Quantum error correction is no longer waiting for hardware. It is waiting for algorithms — and that, finally, is the right order of operations.

Somewhere in the constellation Hydra, about 7.43 light-years from the Sun, an object slightly larger than Jupiter and roughly twice Jupiter’s mass sits at 265 kelvins and slowly rotates, surrounded by water-ice clouds. It is the coldest known brown dwarf, a failed star with no fusion at its core, and on a single clear orbit of the James Webb Space Telescope, a team led by Brittany Miles of the University of Arizona’s Steward Observatory pointed the Near-Infrared Spectrograph at it for eleven straight hours. The result, announced on 8 October 2026 and accepted for publication in The Astrophysical Journal, is the first direct evidence that water clouds thicken and thin on a body outside the solar system. The air is moving. The weather is real. We can now watch it.

Brown dwarfs have always been a category problem. Too massive to be planets — at least thirteen Jupiter masses is the conventional lower bound — and too small to clear their neighborhoods, they are also too small to ignite steady hydrogen fusion. They glow with the residual heat of their formation, dimming visibly year over year as they radiate it away. A young T-type brown dwarf can sit at 1,500 K and look vaguely like a hot Jupiter. A middle-aged one in the L-T-Y sequence cools through a thousand kelvins and changes chemistry, with silicate and iron clouds raining out of its atmosphere at successive temperature plateaus. By the time an object reaches the Y class, somewhere below 500 K, water has frozen into clouds at the level of the photosphere, and the chemistry has become alien enough that it overlaps with what a cold, isolated gas-giant planet would look like. WISE J085510.83-071442.5, the object that Kevin Luhman first identified in the WISE survey in 2013, sits so far down the cooling track that its own temperature is below the freezing point of water. That is why astronomers care about it. It is the closest analog to a free-floating giant planet that we can see in detail (Luhman 2014, ApJ 786, L18).

The new work leans on that proximity. With a parallax of 439.0 ± 2.4 milliarcseconds measured against background quasars (Luhman & Esplin 2014), WISE 0855 is the fourth-closest stellar system to the Sun, after Alpha Centauri, Barnard’s Star, and Luhman 16. At 2.278 parsecs it is close enough that JWST can collect a medium-resolution spectrum, R ≈ 1,000, with high signal-to-noise in fifteen-minute slices across an entire rotation. The team obtained 32 such spectra, one every fifteen minutes for eleven hours, watching the same atmospheric column long enough to track features as they came in and out of view. The result is not a static spectrum. It is a movie with chemical resolution.

What they saw falls into two distinct layers. In the upper atmosphere, between roughly 1 and 4 bar of pressure, the water-ice cloud opacity oscillates. Some rotation phases show deeper, thicker clouds; others show clearer air, with the underlying gas-phase water absorption features rising into view. Below that, near 70 bar, the abundances of carbon monoxide and, in particular, phosphine (PH₃) vary in a way that does not match a quiet atmosphere. The team interprets the deeper signal as convective dredging: parcels of gas from well below the photosphere are rising, cooling, and being mixed into the visible layers, carrying with them the products of high-pressure chemistry. Phosphine, in particular, is hard to make in a cold, hydrogen-rich atmosphere at equilibrium; seeing it vary is a fingerprint of disequilibrium mixing, the same vertical transport that drives Earth’s weather but in a gas made mostly of molecular hydrogen and helium.

Why this matters is partly a question of what the older instruments could not do. Pre-JWST photometry of WISE 0855 measured the total brightness changes as the object rotated, but photometry collapses everything — clouds, chemistry, temperature — into a single flux number. NIRSpec at R ≈ 1,000 resolves individual molecular lines, so a cloud’s signature can be told apart from a chemistry change by the wavelength dependence. A water-ice cloud dims more in the continuum than in a line core; phosphine varies in a sharp, narrow band near 4.3 μm. That wavelength-resolved time series is what turns “the brightness is changing” into “the air is moving,” and only JWST has the combination of light grasp, wavelength coverage, and stability to make the measurement.

The broader payoff is the bridge the result builds between solar-system science and exoplanet science. Jupiter, the only other place where we can stare at water-ice clouds for a full rotation, has its own meteorology, and the cloud structures are different in kind — banded, with sharp shear zones and lightning-driven convection. The brown-dwarf version is slower, colder, and chemically simpler, but the physics overlaps. Carbon monoxide and phosphine chemistry in hydrogen-rich air follows the same quench kinetics, the same vertical-mixing timescales, the same temperature-dependent pathways. Modeling tools built for Jupiter’s atmosphere can be ported across. As the team puts it, the cloud-thickness modulation they see in WISE 0855 sits at pressures close to the photosphere, while the CO and PH₃ modulation sits much deeper; the two signals decouple, which they would not if a single effect were driving both. The weather is layered, the way weather here is layered.

There is a second, quieter implication. Brown dwarfs are not the only objects at this effective temperature. The directly imaged giant planets around HR 8799, β Pictoris, and a handful of other young A-stars sit at L-to-T spectral types, with brightness temperatures around 1,000 to 1,500 K. The colder, free-floating Y-dwarfs like WISE 0855 are the only objects old enough and cool enough to make the photospheric chemistry fully settled, and that makes them the closest natural laboratory for what an isolated Jupiter-mass planet would look like after it stops glowing from formation heat. The same NIRSpec time-series technique, applied to a young, hot Jupiter, will not work — the contrast against the host star is too steep — but applied to a Y-dwarf, it works because the dwarf is its own light source. If the team can repeat the measurement on a slightly warmer or slightly cooler Y-dwarf, they can map how the cloud layer’s vertical position and the convective mixing depth depend on effective temperature. That is a phase diagram of substellar weather waiting to be filled in.

For the moment, the headline is small and concrete: on one eleven-hour night, JWST watched water clouds form and dissipate on an object only 7.43 light-years away. The first exoplanet weather report, in the strict sense, is in. The next step is repetition across the Y class, where a small population of cold brown dwarfs within 20 parsecs is already known and the JWST time-allocation committee is more than willing to spend hours on them. The fields that should move fastest are the cloud microphysics of water ice in hydrogen-rich gas and the rate at which disequilibrium chemistry can reshuffle carbon and phosphorus through an atmosphere with no stellar irradiation to drive it. Both questions have been waiting, in a sense, for a Webb that could simply sit on one cold dot for half a night and watch. That wait is over.

In the last week of September, NASA’s X-59 Quesst flew six times in four days. Two flights on September 22, then two more on the 24th, then two more on the 25th. The previous best was four, so the team didn’t just break its own record; it lapped it. One of those flights also pushed the aircraft to Mach 1.53, roughly 1,010 miles per hour, the fastest the X-59 has gone in the year since Lockheed Martin’s Skunk Works rolled it out of the Palmdale paint shop (NASA Quesst blog, October 2 2026). The run brought the program’s total to 38 test flights, and the aircraft is now in a scheduled three-week maintenance stand-down before the next phase of envelope expansion.

The X-59 is, on paper, the most unassuming supersonic jet ever built. It has no forward windshield. It is shaped like a long black arrow with a 33-foot nose spike, the cockpit pushed so far back that the pilot flies by reference to an external vision system, a four-camera array feeding a single high-resolution display. The whole point of the thing is what it does not do: it is designed so that the shock waves it generates at cruise do not coalesce into a single audible boom when they reach the ground. NASA wants the on-the-ground signature to come in around 75 effective perceived noise decibels, a level the agency compares to a car door closing two houses down (NASA, Quesst: The Vehicle; Lockheed Martin X-59 product page).

The reason that number matters has very little to do with engineering aesthetics and everything to do with regulation. Civil supersonic flight over land is, with limited exceptions, banned in the United States under FAA rules that trace directly back to the sonic boom problem. The Concorde, which flew at Mach 2.02 over the ocean and on a few chosen overland corridors at subsonic speeds, retired in 2003 partly because its boom made overland expansion politically impossible. Twenty-three years later, no commercial airliner has flown faster than sound over the contiguous United States. The X-59 is NASA’s bet that a different shape can produce a different shock structure, and that if the shock structure is quiet enough, regulators will eventually rewrite the rule.

The hardware story begins in February 2016, when NASA awarded Lockheed Martin a preliminary design contract under the Low-Boom Flight Demonstrator program. In April 2018, the agency converted that study into a $247.5 million build contract, the largest single X-plane procurement in a generation. Skunk Works in Palmdale, California, did the structural work. NASA Langley ran an 8%-scale model through three weeks of wind-tunnel testing in late 2017, sweeping angle of attack out to 50° and 88° at low speed to map the static margin and dynamic response of the unusually long forebody. The critical design review closed in September 2019, construction crossed the halfway line in late 2020, and the aircraft was rolled out, fully painted, in January 2024 (Lockheed Martin X-59 Quesst, Wikipedia; NASA Quesst gallery).

The first flight came on October 28 2025, subsonic, out of Armstrong Flight Research Center at Edwards Air Force Base in the Mojave. The aircraft stayed subsonic through the early envelope, and on April 3 2026 the team flew wheels-up for the first time, reaching 20,000 feet and roughly 460 mph. The first supersonic run came on June 5 2026, and a week later, on June 12, the X-59 reached its design cruise point of Mach 1.4 at 55,000 feet. What the late-September record set adds on top of those milestones is volume: not a single new speed or altitude, but the demonstrated ability to turn an experimental one-off into something that flies like a working test article, twice a day, on back-to-back days.

The underlying physics the aircraft is built to exploit is called Mach cutoff shaping, and it is older than the X-59 by several decades. A supersonic aircraft generates two main shock waves, one from the nose, one from the tail, and in a conventional shape those shocks hit the ground as two closely spaced impulses that the ear fuses into a single boom. The trick is to design the airframe so that the upper-frequency components of the pressure signature are refracted upward by the natural temperature gradient of the lower atmosphere, never reaching the surface. The rule of thumb is that a ray of the shock will be turned away from the ground if the local sound speed increases faster than the local Mach number falls as the ray climbs. In practice, that means the X-59 has to be flown within a tight band of altitude and speed; outside that band, the rays bend back down and the boom returns.

The airframe itself encodes the same idea geometrically. The 33-foot nose is shaped to spread the nose shock over a longer distance, lowering its peak overpressure. The planform is built around a long, thin fuselage with a single low-sweep wing and a horizontal tail deliberately set high enough to keep the tail shock from merging with the wing shock. Wind-tunnel Schlieren imaging in 2018 confirmed the resulting pressure signature on the ground: a rise time of about 6 milliseconds and a peak overpressure of about 0.5 pounds per square foot, low enough that listeners describe the event as a thump rather than a bang (Sonic boom, Wikipedia, Mach cutoff references; CNN, August 6 2026).

The reason nobody has been able to demonstrate this in commercial service yet is the public response. A 75 EPNdB thump is in the same A-weighted decibel range as a dishwasher running in the next room, but supersonic cruise would put that thump over populated areas that have never heard anything louder than a turboprop. To find out what people actually think, NASA plans community overflight tests in 2026 and 2027, dropping F/A-18 Hornets into shallow dives from 50,000 feet to reproduce the X-59’s signature without the X-59, then asking residents to rate the noise. The agency needs that data to take to the International Civil Aviation Organization’s Committee on Aviation Environmental Protection, which sets the supersonic noise standard, and to the FAA, which would have to amend 14 CFR Part 91 to permit overland cruise. Without that amendment, an aircraft that thumps instead of booms is still illegal to fly over Denver.

The next three weeks of maintenance, after the record run, are scheduled around exactly that preparation. When the X-59 returns to the air, the team plans to expand the flight envelope further before the first community-response flights. Whether the data ultimately change ICAO and FAA rules is not a question the engineers can answer; they can only build the airframe that makes the question worth asking. The September record tells you how far that build has come: from a concept on a Lockheed whiteboard a decade ago, to an aircraft that can fly two missions a day for three days in a row, faster than any civil airplane has flown over the United States in a generation.

On Sunday morning Japan time, a 4,000-kilogram spacecraft will leave Tanegashima Space Center on Japan’s new H3 rocket and begin a five-year journey to do something nobody has done before: touch a moon of another planet and bring a piece of it home. The mission is the Martian Moons eXploration probe (MMX), and its destination is Phobos, a 22-kilometer rock that swings around Mars three times a day and is slowly falling inward. A small capsule will land in an Australian desert in 2031 carrying at least ten grams of alien dirt that may finally settle a sixty-year argument about where Phobos came from.

That argument is unusually clean for planetary science. Two possibilities fit the observations, and they point in opposite directions. Either Phobos and its smaller sibling Deimos are captured asteroids — dark, carbon-rich bodies pulled in from the outer main belt — in which case they should resemble primitive outer-belt material. Or they are the wreckage of a giant impact on early Mars, a process analogous to how Earth’s own Moon formed, in which case they should be dry, glass-rich, and chemically indistinguishable from Martian basalt. Existing remote sensing cannot break the tie. Phobos is too small and too dark for Earth-based spectroscopy to read its surface cleanly, and the two theories predict mineralogies that differ in subtle ways. Only a sample under a laboratory bench can decide.

That is what JAXA is sending MMX to do. Approved in February 2020 under project manager Yasuhiro Kawakatsu and shipped to Tanegashima in March 2026, the probe lifts off at 04:41 JST on 20 October 2026, 19:41 UTC on the 19th (ESA, 13 October 2026). The probe is built in three modules: a 1,800-kilogram propulsion section carrying about 1,900 kilograms of bipropellant, a 150-kilogram exploration module, and a 1,050-kilogram return module that will eventually fire its own engine, depart Phobos, and re-enter Earth’s atmosphere. The rocket is the H3-24L, the four-solid-booster variant of Japan’s flagship. The H3 failed its first two flights in 2023, and getting it back to operational status was the engineering prerequisite that made MMX possible at all.

Seven instruments ride on the main spacecraft, and four more sit on a small rover called IDEFIX that will be deployed in early 2029. NASA’s contribution is MEGANE, Japanese for “eyeglasses,” a gamma-ray and neutron spectrometer built at the Johns Hopkins Applied Physics Laboratory that reads elements like hydrogen, which cameras cannot see, to determine how much water-bearing material is present. CNES in France built MIRS, a near-infrared spectrometer for mineral identification. JAXA added a LIDAR, a dust counter to look for a hypothesized faint ring of debris around Mars, and a mass spectrometer for the local ion environment. The IDEFIX rover is a joint CNES-DLR build, weighs 25 kilograms on Earth, and will weigh the equivalent of fifteen grams on Phobos (ESA, 13 October 2026).

The path MMX will fly is itself an engineering problem. Phobos is too small for a true orbit — a satellite around the moon would be perturbed away by Mars within days. Instead, MMX will use a quasi-satellite trajectory, a Mars-centered path that lets the spacecraft stay near the moon for a full Martian year. The geometry is governed by the mass ratio between the two, about 1.7 × 10⁻⁷, which is too small to permit a stable bounded orbit around the moon. The same physics that holds objects at the L1 and L2 Lagrange points lets a spacecraft at the right distance and phase angle drift along with the moon indefinitely.

The sampling system is a second engineering story. JAXA is flying two redundant mechanisms. The first is a coring-and-shovel arm of traditional design. The second, called P-sampler, was developed by Honeybee Robotics and is the descendant of a Planetary Society-funded prototype called PlanetVac. Where most regolith samplers depend on a lander’s weight for a percussive motion, P-sampler fires pulses of nitrogen gas to stir the surface into a vortex, then sucks the cloud into a collection chamber, working whether the lander is upright, tilted, or on one footpad. PlanetVac’s path to Phobos is a parable of compounding: a 2013 lab test, a 2018 flight on Masten’s Xodiac rocket in the Mojave, a 2025 demonstration on Firefly’s Blue Ghost lunar lander, and now inclusion on MMX as a primary sample-acquisition method (Planetary Society, 14 October 2026).

The mission timeline runs five years. MMX will arrive at Mars in August 2027, enter a quasi-satellite orbit around Phobos in September, study the moon for roughly eighteen months, release IDEFIX in early 2029, and land in 2030 to collect the sample. The return module will then push off Phobos, swing past Deimos, and head home. The sample capsule is designed to land in the South Australian outback.

What MMX finds is the test. A captured-asteroid Phobos should have the spectral fingerprints of carbonaceous chondrites, with hydrated silicates and organic compounds. An impact-ejecta Phobos should look like a re-melted piece of Martian crust, dominated by olivine and pyroxene in a glassy matrix. Either answer changes planetary science. The capture scenario would re-open the question of how many small moons are interlopers rather than natives. The impact scenario would mean Mars once had a ring system that briefly rivaled Saturn’s. The samples also carry information no remote observation can read: cosmic-ray exposure age, solar-wind noble gases, the isotopic ratios that fingerprint Martian versus asteroidal origin. ESA project scientist Elliot Sefton-Nash put it simply: samples returned to Earth are “the gift that keeps on giving,” because the same ten grams will be analyzed for decades by instruments not yet invented.

That’s why Japan built this spacecraft, and why a rocket that failed its first two flights is about to carry it off the pad. The question of where Phobos came from has been the cleanest unsolved problem in solar-system science for half a century, and MMX is the cleanest possible test of it. Sunday’s launch opens the clock.

For decades, the appeal of free-space optical communications from orbit has been obvious. A near-infrared laser beam carries orders of magnitude more spectrum than a comparable radio antenna, the beam is narrow enough to be effectively unjammable, and the photons cost almost nothing to produce. The catch has always been the last 50 kilometres of air. Index-of-refraction turbulence in the lower atmosphere bends the beam by microradians on millisecond timescales, smearing what should be a clean signal into a dancing speckle pattern on the receiver. Satellites have been talking to each other with optical links for years, but the moment a beam has to punch through clouds, fog, and temperature inversion layers, the system has to be either very large or staffed by a small army of optical engineers. On October 8, 2026, a Lithuanian startup called Astrolight, working with two Greek universities and the European Space Agency, showed that a 6U cubesat, a remote-controlled mountaintop terminal, and a piece of software that does not need a human to babysit it can do the job anyway, at 1 Gbps, with link margins to spare.

The piece of hardware in question is ATLAS-1, a laser communication terminal that Astrolight bolted to a nanosatellite called ERMIS-3, which the company says was deployed on a SpaceX rideshare launch (SatNews, October 8 2026). ERMIS-3 is funded through the Greek national space programme and the European Recovery Fund “Greece 2.0” initiative, with the National and Kapodistrian University of Athens running the Department of Aerospace side of the experiment and the Aristotle University of Thessaloniki operating the upgraded optical ground station. Astrolight’s job was the optics: the transmitter on the satellite, the receiver and beacon laser on the ground, and the software that joins them into a closed feedback loop. The ground station sits on Holomondas, a former astronomical observatory in northern Greece that Astrolight retrofitted earlier in 2026 with an 808-nanometre uplink beacon and a C-band optical receiver, and rebranded as OGS-2.5 (Intelligent CIO Europe, May 20 2026).

What the team reported on October 8 is more interesting than “it worked.” The link achieved first light on the satellite’s initial orbital pass over the ground station, transitioned to routine 1 Gbps downlinks within a week, and now runs two pre-scheduled passes per day, one in daylight and one at night, with no on-site operator. Closed-loop tracking held the beam down to elevations below 10 degrees; for passes above 15 degrees, link availability cleared 85%. The terminal re-acquired an interrupted signal in under 10 milliseconds, the speed the U.S. Space Development Agency now demands for inter-satellite optical links (SatNews, October 8 2026). That reacquisition budget is the operational number that matters. A low-Earth-orbit spacecraft flies past any single ground station in seven or eight minutes, and a slow reacquisition means lost megabits. Ten milliseconds is fast enough to hand the link off cleanly between stations on a single pass, which is what the commercial constellations are paying for.

The reason this is different from the inter-satellite optical links that SDA, Mynaric, and Tesat have been flying for several years is that those terminals never have to deal with the atmosphere. A beam from one satellite to another crosses nothing but vacuum, so the engineering reduces to narrow-beam pointing and high-fidelity wavefronts. A downlink to a mountaintop, by contrast, has to survive turbulence cells roughly 10 centimetres across near the ground, scintillation that flips the received power by a factor of three on second-long timescales, and weather. Holomondas sits at roughly 1,000 metres of elevation, which buys it less air column and fewer low-altitude turbulence cells than a sea-level site would. The 808 nm beacon Astrolight installed at the ground station gives the satellite’s coarse-pointing system a bright target to chase before the fine tracker takes over, and the C-band receiver behind it tolerates the same 1550 nm family of wavelengths that terrestrial fibre networks use, which means the same indium-gallium-arsenide photodetectors and erbium-doped fibre amplifiers can be repurposed from telecom inventory (Intelligent CIO Europe, May 20 2026).

The closed-loop trick is what closes the engineering gap. The beacon tells the satellite where the ground station is in inertial space; the satellite’s fast-steering mirror narrows the beam to a few hundred microradians, narrow enough that diffraction alone would put a metre-scale spot on the ground; the ground station’s wide-field detector sees the spot, measures its centroid, and pushes the offset back up the beacon. The whole loop runs at a few hundred hertz, well above the bandwidth of the turbulence that drives the worst scintillation. Once the loop is locked, the received power stabilises, and the link budget, which Astrolight quotes at 1 Gbps nominal and “up to 10 Gbps” headroom for the ATLAS-1 design, becomes a question of how much laser power the satellite can afford to spend per bit. A useful scaling relation is that the required transmit power scales as the square of the range, so a low-Earth-orbit link at 500 km needs roughly 10,000 times less power than a comparable geostationary link at 36,000 km — which is why nobody has ever put a useful optical downlink on a GEO satellite without a relay mirror in between.

The team has not talked publicly about the exact laser power, the modulation format, or the forward-error-correction scheme. Those numbers matter: at 1550 nm, the atmospheric transmission window is wider than at 1064 nm, and the eye-safety limit allows a few hundred milliwatts of average power before a ground telescope becomes a retinal hazard. The published numbers point toward on-off keying rather than the more exotic quadrature-amplitude-modulation formats that ground fibre uses, because residual scintillation still costs several decibels of margin if you try to ride a higher-order constellation through it. The “up to 10 Gbps” claim is therefore best read as a stretch goal in good seeing, not a guaranteed nightly rate.

What changes because of this is the small end of the market. The same architecture — a 1U-to-6U cubesat payload, a 30 centimetre ground receiver, and a software stack that runs itself — fits Earth-observation operators who today are stuck at tens of megabits per second over X-band because their satellite is too small for a steerable parabolic dish. Astrolight’s roadmap includes the ATLAS-2 and ATLAS-X-MINI terminals at 2.5 Gbps, and the larger ATLAS-X destined for ESA’s HydRON multi-orbit optical network in 2028 to 2029 (SatNews, October 8 2026). The company is also building a second optical ground station in Greenland with ESA, partly to chase high-elevation polar passes that Mediterranean sites cannot see, and partly to use the long Arctic nights as natural immunity against daytime turbulence. The first trans-Atlantic optical link from a commercial LEO satellite is no longer a 2030 problem.

The interesting historical echo is that this is, finally, what NASA and ESA have been quietly trying to build since the 1990s. Japan’s KIRARI satellite demonstrated a 1.5 Mbps optical downlink in 2006. The Lunar Laser Communication Demonstration returned 622 Mbps from the Moon in 2013. The European Data Relay System started operational optical service in 2016 at 1.8 Gbps between geostationary relay and LEO. The difference now is that a 6U cubesat, operated by a small team in Vilnius, can do the job, unattended, and the same company’s roadmap reaches deep into ESA’s flagship programmes. The orbit-to-ground bottleneck has been broken at the small end. What comes next is the same trick repeated a thousand times, every time a constellation operator looks at a 1 Gbps optical downlink and realises it no longer needs a ground crew to make it work.