When most people picture space robots, they imagine wheeled rovers trundling across dusty red terrain. NASA has a different vision for the next era of lunar exploration: humanoid robots, standing upright, working alongside astronauts, handling the jobs that are too dangerous, too repetitive, or simply too time-consuming for humans to do themselves.
At Johnson Space Center in Houston, a team of engineers is quietly building toward exactly that future.
Not a Replacement — A Partnership
The framing from NASA’s Dexterous Robotics Team is deliberate and worth taking seriously. This is not about replacing astronauts. The goal, as team lead Shaun Azimi puts it, is to make human exploration “safer and more sustainable” — not to make humans redundant.
The logic is straightforward. As missions push further from Earth, toward a permanent lunar presence and eventually Mars, the risks multiply and the cost of human error or injury grows. A robot that can handle equipment transport, routine maintenance, dangerous repairs, or the physical grunt work of setting up a base doesn’t take the human’s place. It frees the human to focus on the things that only humans can do well — scientific judgment, exploration, complex decision-making, the unpredictable moments that no algorithm anticipates.
“If we can send more capable robots, we can reduce the risk and make people more effective in doing the things that people do best,” said Azimi.
A Decade of Groundwork
NASA hasn’t arrived at this point overnight. The agency has been developing and refining bipedal robots for over fifteen years. Valkyrie, NASA’s first fully upright humanoid robot, was completed back in 2013. Robonaut 2 — which has already visited the International Space Station — was the agency’s first two-legged robot to reach orbit.
These early prototypes laid the mechanical and software groundwork. What’s being built now is the next generation: robots capable of operating reliably in the specific harshness of the lunar environment. The moon is not a friendly place for machines. Microgravity affects movement in ways that Earth testing can only partially simulate. Lunar dust is extraordinarily fine and abrasive, working its way into joints and mechanisms that would survive indefinitely in a lab. Solar radiation levels are punishing. Engineering a robot to handle all of this dependably, not just for a demonstration but for sustained real-world use, is considerably harder than building one that performs well in a controlled environment.
Testing Before Flying
To bridge that gap, NASA has built the Integrated Mobile Evaluation Testbed for Robotics Operations — iMETRO — at Johnson Space Center. It’s a facility designed to let both NASA teams and private companies test and refine technology under conditions as close to lunar reality as possible, including mock lunar surface environments, living quarters, rover systems, and a rocky outdoor enclosure built to mimic the moon’s terrain.
The value of iMETRO is in catching problems before they happen in space — including problems nobody thought to anticipate. Space missions don’t come with a repair crew and a parts warehouse. Failures that would be minor inconveniences on Earth become serious crises when you’re 238,000 miles away with no immediate way home.
The Small Tasks That Matter Most
One recent project gives a sense of how this work actually unfolds. A company called PickNik tested software directing a robotic arm to identify a spacecraft hatch, open it, and transport cargo bags from one location to another. It sounds almost mundane. But tasks like this — repeated dozens of times during a mission, physically demanding and carrying a constant small risk of injury — are exactly where robotic assistance delivers the most value.
Reducing the cumulative wear on astronauts during routine operations isn’t a minor benefit. It’s the difference between a crew that arrives at its most critical scientific moments fresh and capable, and one that’s already spent significant physical reserves on logistics.
The Broader Vision
The humanoid form factor matters here in a way it doesn’t in many other robotics contexts. The moon base of the future will be built around human-scale spaces, human-scale tools, and human-scale tasks. A robot that can stand, reach, carry, open doors, and manipulate the same equipment designed for human hands doesn’t require an entirely separate infrastructure. It works in the world that’s already been built.
What NASA is developing at Johnson Space Center is not science fiction. Valkyrie is already over a decade old. The question is no longer whether humanoid robots can work in space — it’s how capable, reliable, and trustworthy they can be made before the next generation of lunar missions needs them.
The answer to that question is being built right now, in a lab in Houston, one test at a time.