Google Public Sector and NASA disclosed on November 2, 2026, that they are developing an AI-powered robotic assistant designed to support astronaut health care on future crewed missions to Mars. The system builds on NASA's Crew Medical Officer Digital Assistant, which has been evaluated since 2025, and combines Google DeepMind's physics simulation tools with an open-source robotics model from Physical Intelligence.
Key takeaways
- The robot retrieves medical equipment, assists with health monitoring, transports cargo and organizes tools under voice command and human supervision.
- In lab testing, it located and moved a medical bag successfully 85% of the time.
- It runs on the Pi0 open-source model via Google Cloud, with spatial-awareness technology adapted from JPL's satellite systems.
- NASA's Human Research Program, Johnson Space Center and Jet Propulsion Laboratory are all involved alongside Google DeepMind.
- The project targets Mars missions, where round-trip communication delays with Earth can reach 44 minutes.
The 44-minute figure is the whole point of the project. On the International Space Station or during Moon missions, astronauts can radio Earth and get guidance back in seconds. On a Mars transit, that's not an option — a medical emergency has to be handled with whatever expertise and equipment is already on board. NASA's working theory is that an autonomous assistant capable of locating supplies and tracking vitals without step-by-step ground instructions reduces how much hinges on a crew member's own training in a crisis.
Why a cargo robot and a medical robot are now the same project
What's notable is how unglamorous most of the robot's job is. Organizing tools, managing spacecraft logistics and transporting cargo make up as much of its workload as anything resembling clinical care. That reflects a broader trend in robotics: health-support systems increasingly succeed not by replacing a doctor's judgment, but by handling the physical logistics around care — finding the right bag, carrying the right tray, tracking where equipment went. Hospital logistics robots on Earth have followed the same playbook, and NASA is effectively importing it into a spacecraft.
The 85% success rate for a single task — locating and moving a medical bag — is a modest number for an operational system, but unremarkable for a proof-of-concept stage. Google Public Sector Vice President Jim Kelly framed the motivation plainly: "An astronaut on a future deep space mission to Mars could face severe operational challenges, including round-trip communication latencies of up to 44 minutes."
What's next
The robot is still in proof-of-concept testing, not flight hardware. It runs on roughly 15 edge-computing devices chosen for power efficiency, a constraint that matters more on a spacecraft than almost anywhere else a robot operates. Future testing will likely focus on raising reliability on core tasks before NASA considers it for Artemis-era missions, where every kilogram and watt of power is budgeted years in advance. Source: Digital Health News.