For the first time, a robot has built part of a structure in space without astronauts assembling it, proving a technology that could eventually allow spacecraft, antennas, and enormous space telescopes to be built directly in orbit rather than launched from Earth as a single piece.
Developed by researchers at the California Institute of Technology (Caltech) and funded by the Defense Advanced Research Projects Agency (DARPA), the experimental robotic system successfully assembled a section of a lightweight antenna structure using individual components carried into orbit aboard a SpaceX rocket.
The achievement constitutes a significant milestone in DARPA’s Novel Orbital and Moon Manufacturing, Materials and Mass-efficient Design (NOM4D) program, which aims to develop technologies prepared to construct structures hundreds of meters, or potentially even kilometers, across.
“NOM4D’s objective is to enable efficient, scalable construction of very large structures in space,” DARPA program manager Andrew Detor said in a statement. “By ‘very large,’ I mean hundreds of meters or kilometers in scale. Nothing approaching that scale can be achieved today.”
A Robot That Builds Spacecraft Similar to Construction Toys
The experimental system, called Autonomous In-Space Tensioned Reflector Assembly (ASTRA), launched into low-Earth orbit on March 30, 2026, aboard SpaceX’s Transporter-16 mission.
Roughly the size of a large household refrigerator, the payload carried a robotic manipulator, composite fiber tubes, aluminum joints, and thin reflective material intended to form a 4.6-foot-wide antenna reflector.
Unlike conventional satellites, which are assembled on Earth and folded to fit inside a rocket, ASTRA was designed to construct its structure one section at a time after reaching orbit.
The robot retrieves individual components from onboard dispensers, arranges them, and connects them into a fixed framework. It also attaches a lightweight reflective film before extending each completed section outward.
DARPA compares the approach to assembling structures using K’Nex or Tinkertoy construction sets.
The underlying concept is fairly straightforward. Instead of launching a fully assembled satellite or telescope, future spacecraft could carry compact packages of structural components that robots assemble after reaching their destination.
Such an approach could eliminate one of the most persistent limitations in spacecraft engineering: the diameter of a rocket’s protective payload fairing.
The Robot Succeeded, but the Mission Revealed Major Challenges
Although ASTRA demonstrated that automated construction is possible in orbit, the experiment also exposed the difficulties of building structures in space.
The payload launched with two sections of its framework already assembled. Once in orbit, the robot successfully constructed a third section, demonstrating the assembly process.
However, completing the antenna’s supporting framework required 11 sections, and researchers encountered technical problems that slowed progress.
According to Caltech, the mission is unlikely to finish constructing the entire reflector before operations conclude in December.
Engineers also faced an unusually restrictive troubleshooting environment. The team had only seven or eight communication opportunities each day, lasting approximately 15 minutes each, to transmit commands and examine photographs of the robot’s progress.
When something went wrong, researchers attempted to reproduce the problem using an identical system in their laboratory before sending corrective instructions into orbit.
Although the robot performed the physical assembly operations automatically, engineers on Earth still supervised the process, evaluated its progress, and intervened when unexpected problems occurred.
Nevertheless, successfully constructing a new structural section in orbit demonstrated that the underlying assembly method could work outside laboratory conditions, even if significant engineering difficulties remain before it can be used to build much larger spacecraft.
Building Structures Too Large to Launch From Earth
The potential applications go far beyond the experimental antenna. DARPA envisions robotic construction systems eventually assembling enormous radio antennas, space-based solar power stations, orbital infrastructure, and structures supporting future human expeditions.
For astronomy, the technology could enable telescope reflectors much larger than those that can currently fit inside launch vehicles. Meanwhile, for communications, larger antennas could provide capabilities that would be impractical using conventional deployable spacecraft.
The technology could also support orbital solar arrays designed to collect sunlight and potentially transmit energy elsewhere.
Importantly, ASTRA did not demonstrate construction on those scales. Its achievement was proving that a robotic system could assemble structural components in orbit, an essential step toward developing much larger systems.
DARPA’s broader NOM4D program is also investigating manufacturing materials directly in space. A separate University of Illinois Urbana-Champaign experiment is expected to test a method for producing high-strength composite structures aboard the International Space Station.
For Caltech’s researchers, the immediate priority is improving the construction system’s reliability and simplicity before attempting larger projects.
“Once you get to the size limit where structures cannot be prebuilt on the ground and unfolded, you have to think about how to build them in space,” Caltech aerospace engineering professor Dr. Sergio Pellegrino, who led the ASTRA project, said in a statement. “So, this work could be useful not only for astronomy and astrophysics but also for communications and other areas, such as building very large solar arrays for space solar power.”
Tim McMillan is a retired law enforcement executive, investigative reporter and co-founder of The Debrief. His writing typically focuses on defense, national security, the Intelligence Community and topics related to psychology. You can follow Tim on Twitter: @LtTimMcMillan. Tim can be reached by email: tim@thedebrief.org or through encrypted email: LtTimMcMillan@protonmail.com