Surgeons at the University of California San Diego used two teleoperated Unitree G1 humanoid robots to remove the gallbladders from live pigs, in a preclinical trial published today in Nature. The robots did not act autonomously — human surgeons drove every motion from a control console — but the experiment marks the first time general-purpose humanoid hardware, rather than a dedicated surgical platform, has completed live minimally invasive operations on animals. The G1 starts at $13,500, with dexterous-hand upgrades pushing configurations past $67,000. That compares to $500,000 to several million dollars for Intuitive Surgical's da Vinci Surgical System, the FDA-cleared robot that dominates operating rooms today.
The trial ran two operations. In the first, a human surgeon stood beside the humanoid robot as an assistant. In the second, two teleoperated G1s worked in tandem. Surgeons controlled the robots through a PC-driven console with a stereo headset, using hand motions translated into tool movement and a foot pedal to engage or disengage control. The UC San Diego team, which nicknamed the setup "Surgie," built custom adapters so the humanoid hands could hold standard surgical instruments and wrote software to map intuitive hand gestures to the tool tips.
The pitch is cost and footprint. A da Vinci system weighs about 1,800 pounds and occupies a large fraction of a modern OR. The Unitree G1 stands 5 feet tall, weighs 60 pounds, and can be moved between rooms — or, in principle, between facilities — with ordinary logistics. Shanglei Liu, an assistant professor of surgery at the UC San Diego School of Medicine, framed the case for cheaper, smaller hardware in an interview with UC San Diego Today.
Key facts
- 01UC San Diego surgeons used two teleoperated Unitree G1 humanoid robots to remove gallbladders from live pigs across two minimally invasive operations.
- 02The baseline Unitree G1 starts at $13,500, with dexterous-hand upgrades pushing the cost past $67,000 — versus $500,000 to several million for a da Vinci system.
- 03The G1 stands 5 feet tall and weighs 60 pounds, compared to the 1,800-pound da Vinci Surgical System used in most hospitals today.
- 04The robot's 450mm arm span is well below the 1.6–1.8m reach of an adult human, forcing frequent mid-surgery recalibration.
- 05Teleoperated humanoid latencies run in the hundreds of milliseconds; surgical robotics research targets under 150ms.
The limits showed up during the surgeries themselves. The team paused for several minutes at a time to recalibrate the robots for accuracy or to physically reposition the robot body or arm relative to the instruments. UC San Diego Today reported the operations took much longer than the same procedures performed on specialized surgical systems. The G1's arm span of 450 millimeters — well short of the 1.6 to 1.8 meters typical for an adult human — constrained the working envelope and forced repeated repositioning.
Latency is the other open problem. Current teleoperated humanoid systems generally run in the hundreds of milliseconds of delay between the operator's hand and the robot's response. Prior surgical robotics research has argued for latency below 150 milliseconds to keep fine motor control clinically safe. That gap becomes larger if the surgeon is not in the same building — the deployment scenario, remote care in rural hospitals or on a battlefield, that makes the whole premise interesting.
Benchmark tasks also favored the incumbents. Both new surgical residents and experienced surgeons performed faster on practice work using the da Vinci Research Kit — the standard hardware for telerobotic surgery research — than when driving the humanoid robots. The added cognitive load of monitoring the humanoid's posture and reach appears to be a real cost, not just a training artifact.
The research team is continuing to iterate on the teleoperated system while also targeting autonomy. Michael Yip, a professor of electrical and computing engineering at UC San Diego, described a goal of building an "autonomous surgical assistant" that could handle general OR tasks such as fetching tools or cleaning up between cases, alongside human surgeons.
Autonomy is where humanoid surgical robots run into the same wall as the rest of general-purpose robotics. Most robotics researchers place safe, fully autonomous operation around humans years away, and the tolerance for error in an OR is roughly zero. The near-term path for hardware like the G1 is almost certainly teleoperation with progressively more assistance, not lights-out surgery.
The economic argument is the one worth watching. If a $67,000 humanoid can eventually match a $2 million dedicated surgical robot on a defined set of minimally invasive procedures, the addressable market changes shape entirely — smaller hospitals, ambulatory surgery centers, and clinics in regions that have never been able to justify a da Vinci install become candidates. Intuitive Surgical's moat is not just the hardware; it is regulatory clearance, training pipelines, and a decade-plus of clinical data. A Unitree-based platform has none of that yet, and Nature-published pig trials are a long distance from FDA clearance for human use. But the cost delta is large enough that regulators, hospital systems, and Intuitive's competitors will be paying attention to how quickly the reliability, latency, and reach gaps close.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
