Unitree robots performed an operation and appeared in Nature

A “da Vinci” surgical robot system that costs tens of millions of yuan is too expensive—so is there a “good alternative”?

In medicine, attention is shifting to humanoid robots that can throw punches and dance.

Recently, international top journal Nature published a paper titled “Feasibility study on in vivo surgery using humanoid robots.” The research team chose the UT-AUSTECH G1 humanoid robot, commonly used in university labs. It held laparoscopic instruments, and through remote operation, it performed laparoscopic cholecystectomy on two live pigs.

This is the world’s first case in which a humanoid robot completes a full standard minimally invasive surgery workflow in vivo. Public information shows that the paper’s first author and corresponding author is a “post-00s” Chinese PhD student—Liang Zekai. Liang Zekai graduated with a bachelor’s degree from Huazhong University of Science and Technology in 2023, obtained a master’s degree at the University of California, San Diego in 2025, and is currently pursuing a PhD in the laboratory of Professor Michael C. Yip at the same university.

Looking back at the entire procedure, in the early preparation stage—so that the UT-AUSTECH G1 could pick up surgical instruments—the research team first “modified” it. They designed a custom fixture connected to the G1’s robotic hands, enabling the robot to grip a commercial manual wrist laparoscopic forceps.

Next was building the remote-operation framework. The doctor sat at the control console, wore a stereo high-definition headset to view the endoscopic image, and used both hands to hold two master control joysticks to operate two instruments. The control console communicated with the G1 via network communication; after proportional scaling, the doctor’s movements were mapped to the robot’s wrist motions.

After obtaining the necessary approvals, the team selected two domesticated female pigs aged about 11 weeks and 16 weeks, with anesthesia managed end-to-end by a licensed veterinarian. Before the operation, the team performed sterile preparation and draping on the animals.

They then rolled the UT-AUSTECH G1 humanoid robot to the bedside, aligned it with the surgical area, and used safety belts as a preventive measure. The entire deployment process required close coordination between the surgical team and the engineering team, with repeated adjustments to the robot’s position, instrument alignment, and the motion range of the trocars (a surgical puncture instrument, also called a Trocar, which is a medical consumable used in minimally invasive surgery).

During surgery, a senior surgeon sat at the control console and remotely operated the manual wrist laparoscopic instruments via the master joysticks and the stereo display. Another senior surgeon or a clinical research fellow served as the bedside assistant, responsible for camera control, tissue traction, exposure, as well as instrument adjustments and cleaning the camera.

There were two surgeries in total: in the first case, the humanoid robot was the main operator, with a human surgeon as the bedside assistant. During the procedure, the team also briefly used a second humanoid robot to support camera holding and retraction. In the second case, the humanoid robot again served as the main operator; bedside assistance was still primarily provided by humans, but there was no collaboration between two robots.

The results showed that there were no major intraoperative complications in the first case. In the second case, there was minor bile leakage and bleeding from the liver bed, but it was properly resolved through suction and electrocautery.

In terms of time, the control console time decreased from 56 minutes and 15 seconds in the first case to 31 minutes and 59 seconds in the second case (shortened by 24 minutes and 16 seconds), and the number of robot deployments decreased from 8 times to 4 times.

At the same time, the paper also did not shy away from pointing out many current shortcomings of humanoid robots, including latency issues, limited working space, the need for frequent calibration during surgery, and a lack of components that can be sterilized under high pressure.

Even though the gap is obvious, the research team remains optimistic about the future. The paper mentions that when the early da Vinci robot performed the first laparoscopic cholecystectomy, it took 6 hours, whereas similar surgeries now take only 30 minutes. “With further development, humanoid robots may provide a scalable alternative to traditional robot platforms and expand robotic capabilities beyond conventional operating rooms.”

The starting point of this research is that the healthcare system is facing a severe manpower shortage. The paper states: “Personnel shortages and the increasing demand for nursing are widening the gap between clinical workload and available skilled labor.” Traditional surgical robots like the da Vinci system are excellent in performance, but they are expensive, occupy a large footprint, and require dedicated instruments and operating-room renovations.

By contrast, humanoid robots have natural advantages: they are similar in body size to humans, so they can directly use tools designed by human surgeons and work in existing operating rooms.

For the industry, the value of this paper lies not in the technology itself, but in the industrial imagination space it opens up.

It’s like factory debates about humanoid robots versus automated production lines: sometimes people question why humanoid robots are still needed when mature automated production lines already exist. The answer lies in flexibility and generalization. Production lines are heavy assets—customized for specific products—with extremely high modification costs; while humanoid robots can enter existing operating rooms, use existing instruments, and work within existing workflows. This kind of generality is exactly where humanoid robots’ potential lies.

Source: Shanghai Securities News

Risk warning and disclaimer

Market has risks; investment is需谨慎. This article does not constitute personal investment advice, and it does not consider any specific investment goals, financial situations, or needs of individual users. Users should consider whether any opinions, viewpoints, or conclusions in this article are consistent with their specific circumstances. Invest at your own risk based on this.