The first print failed. In August, one of the four printers on Orbital Matter's Replicator 2 satellite extruded resin into the vacuum and cured it under ultraviolet light. The layers did not bond. Gaps opened between them. A structure that looked solid on camera was a stack of loose rings.
Two months later, on 21 October 2026, a second printer ran for about 30 minutes and produced a continuous boom, 30 centimetres long, with clean bonding between layers. It was the first rigid structure of its kind built in the open vacuum of space, outside the pressurised cabin of the International Space Station. The boom then partially deployed a solar array.
The interesting part is the failure that made the success cheap.
In-space manufacturing moved from a pressurised lab to open space.
The path to large solar arrays became a printing problem, not a folding one.
Why folding sets the ceiling
Every large structure in orbit today is an origami exercise. Solar arrays, radiators and telescope mirrors are folded to fit a rocket fairing, then unfolded through a choreography of hinges and latches. Starlink V2 satellites fly solar arrays about 20 metres long. A single ISS solar wing is roughly 35 metres. Both sit near the practical limit of what a fairing can carry and a deployment mechanism can reliably open.
Orbital Matter inverts the constraint. Instead of designing the spacecraft around the launch, it prints the support structure where it will be used. The binding constraint becomes manufacturing location, not fold geometry. The company's engineers put a 100-metre boom within reach, nearly three times the span of an ISS wing, limited mainly by how much resin the satellite can carry.
What Replicator 2 actually proved
Replicator 2 launched on 7 July 2026 on a SpaceX Transporter-17 rideshare, using an EnduroSat bus. The satellite weighs about 13 kilograms and carries four printers. Each one extrudes photopolymer resin through a precision head, cures it with UV light, and pushes the hardened beam outward on a spinning cylinder. The whole mechanism uses a single motor and draws about 20 watts at peak.
The first print, on 12 August, confirmed the process survives vacuum, temperature swings and microgravity. It did not confirm quality. Between August and October the team ran terrestrial print jobs, tuned parameters and rewrote the firmware. The second print passed. A camera feed drove an image-recognition loop that checked the material as it was laid down, the same closed loop the company needs as booms get longer.
One of the two printers assigned to the solar array failed after an anomaly. The surviving boom still deployed the array. Four printers, one satellite, and a deliberate bet on learning in orbit instead of testing every case on the ground.
The 2028 bet: 5 kW for orbital data centers
The next step is a 2028 mission with EnduroSat to deploy a 5 kW solar array, roughly 5 metres long, on a larger satellite with larger printers. The target customer is explicit: orbital data centers and AI compute in orbit, which need hundreds of kilowatts per satellite.
Building the array's structure in orbit opens up a whole new realm outside of today's deployable structures limit.— Victor Danchev, CTO, EnduroSat
Danchev says the partnership aims to scale to 100+ kW arrays at a fraction of the current market cost per watt. Orbital Matter's roadmap runs from the PADS module — booms from 10 metres, 38 to 90 mm in diameter, now in flight validation with ESA and Thales Alenia Space — to kilometre-scale arrays by 2028 and 100+ kW per wing by 2030.
Seven people and a dead end that mattered
The company was co-founded in 2022 by Jakub Stojek and Robert Ihnatisin, who came from the 3D-printing firm FormLabs. It has raised a $3 million seed round and a €500,000 European Innovation Council pre-accelerator grant, and employs seven people. It is raising again to grow the team and open a printer production line.
The dead end is instructive. NASA funded Redwire to print structural beams in a vacuum chamber in 2023; the spacecraft for that work never flew. The idea of printing in space is old. Orbital Matter's contribution is getting the printer outside and proving the loop closes there — a smaller, duller milestone that happens to be the one that counts. As we wrote in November, Flexell Space bought two Falcon 9 slots to prove solar hardware in orbit; the microgravity-services trade got its own treatment in our Tumbleweed piece. Both point at the same bottleneck: power, built where it is used.
What to watch
The 2028 mission is the real test. A 50-centimetre boom is a proof of process. A 5-metre, 5 kW array is a product. Between them sits unglamorous work: scaling a printer, qualifying a material for a decade in vacuum, and convincing satellite builders that a printed boom is cheaper than a folded one. If that holds, the fairing stops being the ceiling on what goes to orbit.