DARPA says NOM4D completed first automated in-space assembly demonstration

DARPA announced on Nov. 5, 2026, that its NOM4D program completed the first automated in-space assembly demonstration, with a California Institute of Technology payload building a truss structure in low Earth orbit using a gantry robot. cite_list: [”Universe Today”, “darpa.mil”, “Tech Insider”, “ALLATRA Media”, “Shattered”]

The demonstration is called Autonomous In-Space Tensioned Reflector Assembly, or ASTRA, and Sergio Pellegrino leads the ASTRA team at Caltech. “This is the first time that a structure has been autonomously constructed in space,” Pellegrino said. Andrew Detor is the NOM4D program manager at DARPA. cite_list: [”Tech Insider”, “Shattered”]

The robot is a relatively simple gantry able to move along three linear axes—left and right, forward and back, and up and down—along fixed rails. Detor credited the Caltech team with solid engineering rather than complicated robotics or advanced algorithms. He compared the approach to Tinkertoys or K’nex construction toys, in which individual truss members are joined at nodes one section at a time. cite_list: [”Universe Today”, “Heise Online”, “darpa.mil”, “Tech Insider”, “ALLATRA Media”, “Shattered”]

He said each section is built piece by piece, a reflective mesh is attached, and the assembled sections are deployed as they are complete. The gantry joined composite truss members with custom-designed joints and deployed a thin reflective mesh during the same automated process. cite_list: [”Universe Today”, “darpa.mil”]

The intended structure used composite fiber tubes, aluminum joints, and a Kapton reflective surface. The planned ASTRA structure was a reflector antenna 1.4 meters in diameter, but the team built only the initial section because of time limitations, so the complete dish was not assembled. cite_list: [”Tech Insider”, “Shattered”]

The 1.4-meter figure is the planned diameter, not a measured result from orbit; a finished, tested antenna would require radio-frequency measurements and performance data. DARPA’s earlier description said the demonstration would test assembly technology using a surrogate mesh rather than full radio-frequency electronics, so the result proves construction, not a working communications antenna. cite_list: [”Tech Insider”, “Shattered”]

The robot operated without astronauts; no crew member guided a tool, no spacewalk set up the workspace, and no station robotic arm did the heavy lifting. cite_list: [”Tech Insider”, “Shattered”]

ASTRA lifted off on March 30, 2026, from Vandenberg Space Force Base into low Earth orbit aboard SpaceX’s Transporter-16 rideshare mission. It rode on Momentus Inc.’s Vigoride 7 orbital vehicle as one of 10 payloads. Momentus hosted the Caltech payload on its Vigoride platform, assisted with launch integration, and supplied communications, power, imaging, and other services during the demonstration. cite_list: [”darpa.mil”, “Tech Insider”, “ALLATRA Media”, “Shattered”]

The payload was completed, tested, and delivered to Momentus in less than 12 months. The robot worked in space for approximately six months before the result surfaced. cite_list: [”Tech Insider”]

In February 2025, NOM4D changed its timeline, replacing an expected third phase of laboratory tests with on-orbit tests. The team went from the idea of an on-orbit demo to launching a working prototype in 15 months. Detor said the maturity was there and there would be more impact demonstrating capabilities in space rather than making things more precisely in the lab. cite_list: [”Universe Today”, “Heise Online”, “darpa.mil”]

NOM4D was not initially structured around an on-orbit demonstration; after promising Phase 2 results, Detor decided to pivot the final phase to pursue a space demo because the system worked so well on Earth. Andrew Detor said originally Phase 3 was going to be about making things more precisely in the lab than was done in Phase 2. cite_list: [”Universe Today”, “Heise Online”, “darpa.mil”]

To date, space structure assembly is done either completely on Earth, limited by launch vehicle fairing size, or by hand-assembling large components delivered through multiple separate launches. DARPA presents the Caltech system as a step toward hardware that travels as a box of building materials and is assembled where it will operate. cite_list: [”Heise Online”, “darpa.mil”]

Past stations such as the International Space Station were assembled module by module using astronauts and robotic arms over many missions, a crewed model that differs from the small commercial payload used for ASTRA. ASTRA removes the crew and the station, performing automated assembly steps without a human in the loop. cite_list: [”Universe Today”, “Tech Insider”]

Tech Insider and Shattered say ASTRA shifts orbital construction from a government-scale human operation to a rideshare payload on a commercial transfer vehicle, lowering the barrier for research labs and commercial operators. The ISS model required crew time, spacewalk preparation, and a large, expensive facility to host the effort. cite_list: [”Tech Insider”, “Shattered”]

NOM4D stands for Novel Orbital and Moon Manufacturing, Materials and Mass-efficient Design and is pronounced “Nomad.” It is a Defense Advanced Research Projects Agency program, the research agency of the U.S. Department of Defense. DARPA states that NOM4D’s goal is to develop the foundations of building robust, precise structures in space. cite_list: [”Universe Today”, “Heise Online”, “darpa.mil”]

The program aims to enable efficient, scalable construction of very large structures in space, hundreds of meters or kilometers in scale, a scale nothing approaching can be achieved today, according to program manager Andrew Detor. “By ‘very large,’ I mean hundreds of meters or kilometers in scale,” Detor said. cite_list: [”Universe Today”, “Heise Online”, “darpa.mil”, “ALLATRA Media”]

DARPA says automated assembly is ideal for large, low-mass designs, and the benefits vary based on the desired type of structure. Such large structures could include modular data centers in orbit or satellites for Space-Based Solar Power. cite_list: [”Heise Online”, “darpa.mil”]

Andrew Detor said the demonstration uses no complicated robotics or advanced algorithms, calling it solid engineering from the Caltech team that combines operations in a way that scales. He also said operating in space is much harder than on the ground because there is virtually no margin to address issues, and testing in the operational environment is instrumental to rapidly maturing the technology. cite_list: [”Universe Today”, “darpa.mil”]

Andrew Detor said the NOM4D Caltech demonstration is not an endpoint but a beginning, asking what large structures will be assembled thousands of miles above Earth. cite_list: [”Universe Today”, “darpa.mil”]

Later in 2026, the NOM4D team from the University of Illinois Urbana-Champaign will launch a project to the International Space Station to demonstrate composite material processing using a space-compatible resin system for energy-efficient polymerization. The upcoming test will use a monomer resin and an energy-efficient heating method to create a strong polymer in space, similar to Carbon Fiber Reinforced Polymer manufacturing for aerospace components. cite_list: [”Universe Today”, “darpa.mil”, “ALLATRA Media”]

As NOM4D winds down over the next few months, the program aims to help jumpstart ambitious ideas and big investments in space manufacturing and assembly. cite_list: [”darpa.mil”]

Tech Insider and Shattered say thermal swings in orbit, vibration from vehicle thrusters, and the precision of joint alignment all grow harder as a truss gets longer. The gantry trades flexibility for predictability, cannot reach around corners like a seven-joint arm, and repeats the same motion with little drift. cite_list: [”Tech Insider”, “Shattered”]

Shattered says the gantry can only work inside the box its rails define, and that building something much larger than the machine would require either a robot that crawls along its own work or a design that grows the rails as the structure grows. cite_list: [”Shattered”] Shattered says one demonstration finishing an initial section does not show the robot can place hundreds of members with consistent accuracy, and that thermal cycling shifts materials while small errors can stack up across a lattice. cite_list: [”Shattered”]

Shattered says the team ran short of time on this flight, suggesting future missions will need longer operating windows, though whether DARPA or another sponsor books those windows is not stated. cite_list: [”Shattered”]

Shattered says space vehicles have flown for decades with folded panels, booms, and arrays that deploy after arrival, while ASTRA instead used loose parts assembled in orbit. Shattered cites the James Webb Space Telescope as the famous example of a huge mirror that launched folded, and notes that folded designs get harder as size grows and each new fold adds a place for something to snag. cite_list: [”Shattered”]

Tech Insider and Shattered say sharing a ride with nine other payloads is far cheaper than buying a dedicated launch, though the exact price of this flight was not disclosed. cite_list: [”Tech Insider”, “Shattered”] Tech Insider says astronaut time is scarce and expensive, no crew rides in low Earth orbit on a small commercial vehicle like Vigoride, and a robot working without a human in the loop makes assembly viable at places where people will not be present for years. cite_list: [”Tech Insider”]

Tech Insider and Shattered say government interest is often the first step before commercial investment follows, and that satellite makers, antenna suppliers, and in-space servicing startups will read the result as a reason to pitch their own assembly concepts. cite_list: [”Tech Insider”, “Shattered”]

Tech Insider says companies that host or ferry payloads, like Momentus, would gain a new service type if robotic assembly becomes a standard payload, and makers of composite tubes and precision joints would gain a new kind of customer. Shattered says composite tube makers and precision joint machinists get a new customer type, and that parts designed for robot handling could push suppliers toward standardized interfaces. cite_list: [”Tech Insider”, “Shattered”]

Tech Insider says defense and intelligence users are the obvious early buyers since DARPA funds the work, commercial satellite firms would follow if the economics hold, and launch providers face a mixed picture because lighter, denser payload stacks could cut the number of rocket volumes each project needs. cite_list: [”Tech Insider”]

Current rocket fairings cannot fit the largest desired space hardware, so smaller components must launch inside them and be assembled once in orbit. Tech Insider says because assembly happens after arrival, the truss never faces launch loads and can in principle be lighter and larger. Tech Insider says rocket fairings cap the diameter of anything launched whole, and while engineers use origami-style deployable arrays and hinged booms, each fold adds a possible failure point and the largest apertures remain out of reach. A deployment mechanism that fails to open can sink a multi-year mission. cite_list: [”Universe Today”, “Tech Insider”]

Tech Insider says SpaceX’s reusable Starship effort aims to lower the cost of lifting mass and widen usable payload volume, and that orbital assembly and bigger rockets compete in some cases and complement each other since a giant rocket still cannot launch an antenna wider than its own fairing while a robot-built truss can exceed that limit. cite_list: [”Tech Insider”]

Tech Insider contrasts AST SpaceMobile’s approach, which leans on big launchers and big folded arrays, with ASTRA’s bet that the array can arrive as parts. cite_list: [”Tech Insider”]

Tech Insider says NASA’s own lunar plans lean on similar thinking, with delivery getting cargo to the surface and assembly tools like ASTRA potentially turning that cargo into working infrastructure. Tech Insider and Shattered say lunar habitat frames, radio masts, and solar structures face a similar delivery problem, and that NASA is spending heavily on surface infrastructure. cite_list: [”Tech Insider”, “Shattered”]

Universe Today describes the proposed Lunar Gateway as a modular design built in space from pre-made components, intended as a docking station and communications relay for trips to the lunar surface and operations there. Universe Today says robotic assembly could feasibly reinforce the Moon’s subsurface lava tubes and turn them into habitats. cite_list: [”Universe Today”]

Universe Today describes a liquid mirror telescope that could be built in a lunar crater from components shipped individually, and cites NASA’s Lunar Crater Radio Telescope concept for the lunar far side. cite_list: [”Universe Today”] Universe Today reports that the rapid advancement of NOM4D surprised even program personnel, and the successful on-orbit demonstration was not expected to happen so soon. cite_list: [”Universe Today”]

DARPA announced the result as the first-ever demonstration of automated in-space assembly, but that priority claim is the agency’s own, and no independent assessment of it has been published. cite_list: [”ALLATRA Media”]

Shattered says ASTRA wins on cost and risk against crew-assisted assembly because nobody needs to go outside, but loses on flexibility since a human can improvise a fix and a simple gantry cannot. Shattered says ASTRA offers potential scale against folded deployables but gives up the decades of flight experience those designs carry. cite_list: [”Shattered”]