Pasadena, California — November 10, 2026: Researchers at the California Institute of Technology (Caltech) have reached a milestone in space engineering after a robotic system successfully assembled part of a structure in orbit without astronauts physically handling the components. The experiment, known as the Autonomous In-Space Tensioned Reflector Assembly (ASTRA), demonstrates a new way of constructing equipment beyond Earth and could eventually help make much larger telescopes, communication antennas and solar arrays possible.

The end effector of a space shuttle robotic manipulator photographed against Earth's horizon.
A NASA space-shuttle robotic manipulator above Earth. Illustrative image, not the ASTRA robot. Photo: NASA. Public domain.

The achievement was announced on November 5 by the US Defense Advanced Research Projects Agency (DARPA), which funded the work through its Novel Orbital and Moon Manufacturing, Materials and Mass-efficient Design (NOM4D) programme. Caltech confirmed that its robot built an additional section, or bay, of a truss structure in orbit. The demonstration did not complete the full planned antenna, but it provided an important proof of concept for automated orbital construction.

A Historic Step Towards Building Structures In Orbit

For decades, engineers have faced a fundamental limitation when designing large space structures: everything launched from Earth must fit inside a rocket’s payload fairing, survive the forces of launch and then be deployed or assembled in orbit. Large telescopes, antennas and solar arrays are therefore often designed to fold into compact shapes or to be transported in sections.

ASTRA tests a third approach. Instead of launching a fully assembled structure, a spacecraft can carry construction materials and a robotic system that joins the pieces after reaching orbit. The robot retrieves composite-fibre struts and connecting joints from dispensers, positions them and joins them to create repeating structural sections known as bays. Those bays form a larger framework called a truss.

A space station truss segment being transferred between robotic arms during Space Shuttle mission STS-119.
A truss segment is transferred using robotic arms during a NASA shuttle mission. Illustrative context, not ASTRA. Photo: NASA. Public domain.

DARPA describes the approach as similar to assembling a construction toy: individual structural members are connected one section at a time, then extended as the framework grows. The significance is not that a robot has built a giant facility already, but that the basic process of automated assembly has now been demonstrated in the operational environment of space.

How The ASTRA Robot Works

Caltech’s ASTRA system was designed to construct a reflector antenna measuring about 1.4 metres in diameter. Its planned structure uses composite-fibre tubes, aluminium joints and a lightweight reflective surface made from Kapton, a heat-resistant polymer film used in spacecraft applications.

The robotic manipulator moves along three axes inside the assembly system. It collects the structural pieces from nearby dispensers, positions them on an assembly plate and connects them to form a new bay. The system is also designed to attach a thin, perforated reflective film to the supporting framework. The film forms two reflective cable nets held under tension by springs.

The reflective material was developed through collaboration involving Caltech researchers and a team at the University of Pennsylvania led by Professor Igor Bargatin. Caltech says the perforated film can require less pre-tension than conventional knitted metallic mesh, offering a route towards lighter reflector structures.

A robotic arm being inspected during assembly of NASA's Phoenix Mars lander in a clean-room facility.
A robotic arm inspected during spacecraft assembly. Illustrative example, not ASTRA hardware. Photo: NASA/JPL/University of Arizona/Lockheed Martin. Public domain.

When a section is assembled, the design calls for it to be pushed outward, making room for the next section to be built. The complete ASTRA antenna design calls for 11 connected bays. During the orbital demonstration, the robot built the third bay after two sections had already been preassembled before launch.

Six Months In Orbit And The Challenges Of Remote Operations

The ASTRA payload was launched on March 30, 2026, aboard SpaceX’s Transporter-16 mission from Vandenberg Space Force Base in California. It travelled to low Earth orbit on Momentus Inc.’s Vigoride 7 orbital vehicle. The payload was approximately one cubic metre in size and carried the robotic assembly equipment, two prebuilt truss bays and the components needed for further construction.

Once the system reached orbit, the team transmitted instructions to its onboard computer through the Vigoride vehicle. Although the robot performed the physical assembly sequence automatically, engineers on Earth monitored its progress using images sent back from the payload. When unexpected problems occurred, researchers investigated them and adjusted the commands.

Caltech reported that the team had only about seven or eight communication windows on many days, each lasting approximately 15 minutes. To troubleshoot, the researchers used a duplicate flight-spare system in the laboratory to reproduce problems and test possible fixes before sending revised instructions to the orbiting payload.

An astronaut operating the Space Shuttle Discovery's robotic arm in orbit.
Astronaut Joseph Acaba operates Discovery’s robotic arm. ASTRA’s assembly was performed by its onboard robot. Photo: NASA. Public domain.

The difficulties were part of the experiment’s purpose. A mechanism that works reliably in a laboratory can behave differently in orbit, where temperature changes, vibration, limited communication opportunities and the absence of hands-on maintenance create additional constraints. The mission gave engineers practical information about how future orbital construction systems should be designed and tested.

Why The Breakthrough Matters For Space Telescopes

One of the most promising applications is the construction of large space telescopes. Bigger mirrors and reflective surfaces can collect more light, potentially allowing astronomers to study fainter and more distant objects. But very large optical structures are difficult to fit inside rocket fairings, even when designed to unfold after launch.

Robotic assembly could eventually allow telescope frameworks and reflective components to be transported separately and joined in orbit. This could give future missions greater flexibility in the size and configuration of their instruments. It does not mean that giant telescopes can be built immediately; optical alignment, structural stability, servicing and precision manufacturing would still require substantial advances.

Engineers assembling the Hubble Space Telescope before its launch into orbit.
Hubble Space Telescope hardware being assembled on Earth before launch. Photo: NASA Hubble Space Telescope. CC BY 2.0.

Could Robots Build Giant Communication Antennas?

Large radio-frequency antennas are another possible use. Bigger antenna apertures can support specialised communications and scientific observations, including radio astronomy. Building their supporting frameworks in orbit could reduce the constraints imposed by launch-vehicle dimensions and make designs possible that would be difficult to transport as complete units.

DARPA’s NOM4D programme is intended to develop efficient and scalable ways to construct very large structures in space. Programme manager Andrew Detor has described a long-term ambition involving structures hundreds of metres or even kilometres in scale. Those dimensions are future goals, not capabilities demonstrated by the ASTRA mission.

A small satellite deployment system attached to the Japanese module's robotic arm on the International Space Station.
A small-satellite deployment system attached to a robotic arm on the International Space Station. Photo: NASA. Public domain.

The Long-Term Possibility Of Solar Power Stations In Space

Space-based solar power is another ambitious field that could benefit from orbital construction. The concept involves collecting sunlight with large solar arrays in orbit and transmitting energy wirelessly to receiving stations on Earth. Such systems would require extensive structures and highly reliable power-transmission technology.

Caltech’s Sergio Pellegrino is also involved in the university’s Space Solar Power Project. The ASTRA demonstration does not prove that commercial space-based power stations are ready to be built, nor does it solve the difficult questions of cost, conversion efficiency, transmission safety and infrastructure. It addresses one enabling challenge: how large, lightweight structures might be assembled after launch.

A space shuttle robotic manipulator silhouetted against Earth's horizon.
A robotic manipulator above Earth, illustrating the orbital environment in which future space infrastructure could be assembled. Photo: NASA. Public domain.

What The Experiment Did Not Achieve

A NASA robotic arm transfers a truss segment during an earlier space shuttle mission.
Earlier orbital truss handling illustrates the scale of the engineering challenge; it is not ASTRA mission hardware. Photo: NASA. Public domain.

The achievement should be understood as a proof of concept. ASTRA did not finish the complete 1.4-metre reflector antenna, and it did not build a giant telescope, a kilometre-scale antenna or a solar-power station. Caltech said time limitations and technical challenges meant the team completed the initial orbital construction section rather than the full planned structure.

There is also an important distinction between autonomous assembly and a mission operating without human involvement. The robot carried out the physical construction operations automatically, but researchers on Earth sent instructions, monitored images and worked through problems. Future systems will need greater reliability, faster assembly and the ability to cope with a wider range of faults before they can be trusted with complex construction projects.

A New Direction For Space Engineering

A satellite deployment system positioned on a robotic arm at the International Space Station.
Orbital deployment technology provides a glimpse of the infrastructure that future in-space construction could build upon. Photo: NASA. Public domain.

For much of the space age, engineers have designed structures around the limits of launch vehicles, using folding mechanisms, deployable components and multiple launches to get large equipment into orbit. Robotic assembly introduces another option: sending construction materials and machines into space and building the structure where it will operate.

The ASTRA experiment is an early step, but an important one. If the approach becomes more reliable and scalable, future missions could construct larger scientific observatories, communication systems and energy infrastructure beyond the dimensions that can be launched as a single preassembled object.

For now, Caltech and DARPA have demonstrated that a robot can assemble part of a structure in orbit. The next challenge is to make the process simpler, more dependable and capable of handling larger projects. The future of space engineering may increasingly involve not only launching spacecraft from Earth, but also building them after they arrive.

Sources And Further Reading

Image and licensing note: Photographs are illustrative examples of NASA space robotics, satellite deployment and telescope assembly; they are not photographs of the Caltech ASTRA mission. NASA images credited here are public domain in the United States under NASA’s image-use policy. The Hubble assembly photograph is licensed under Creative Commons Attribution 2.0 (CC BY 2.0). Individual credits appear in captions.