Unitree G1 Humanoid Robot performs World's First Live Porcine Cholecystectomy on Nature: 27kg General Platform Challenges Da Vinci

Unitree G1 Humanoid Robot performs World’s First Live Porcine Cholecystectomy on Nature

1. Introduction: When a “Plug-and-Play” General Robot Enters the Operating Room

In July 2026, a UCSD team published in Nature confirming that a 27kg Unitree G1 humanoid robot successfully performed a standard laparoscopic cholecystectomy (gallbladder removal) on a live pig — 56 minutes for single-robot operation, and just 32 minutes for dual-robot cooperative operation without human intervention.

This is the first time a general-purpose humanoid robot has achieved millimeter-precision surgery on living tissue. Unlike the Da Vinci surgical system (priced at $3-4 million, weighing half a ton), the G1 is a commercial off-the-shelf product costing only ~$40,000. This breakthrough proves that low-cost general-purpose robot platforms have crossed the threshold into high-value medical applications.


2. Surgery Details

2.1 Configuration

Configuration Single G1 Dual G1 Cooperative
Robot 1× Unitree G1 (27kg) 2× Unitree G1
End Effector 3D-printed laparoscopic tool Main: laparoscope + cautery; Assistant: grasper + suction
Control Remote (surgeon via haptic手柄) Autonomous cooperative (pre-set protocol)
Subject Live pig (~50kg) Live pig
Duration 56 minutes 32 minutes
Complications None (<5ml blood loss) None

2.2 Critical Steps

Key precision requirements:

  • Calot’s triangle dissection: 0.3mm — the most dangerous step
  • Cystic artery clipping: 0.2mm — mis-clipping could cause massive hemorrhage
  • Gallbladder bed dissection: 15 minutes — requires sustained force control

2.3 Cost Comparison

Metric G1 Humanoid Da Vinci Si
Weight 27kg ~500kg
Price ~$40,000 $2-4M
DOF 43 (full body) ~7 (single arm)
Versatility General purpose Surgery only
Surgery time 56 min ~40 min
10-year total cost (3 G1s) ~$300K ~$5M+

3. Technical Challenges: Where Does “Surgical Precision” Come From?

3.1 Hardware Foundation

G1’s key specs for surgical application:

  • 43 DOF full body, 7 DOF per arm
  • Repeatability: ±0.1mm
  • Force resolution: 0.05N
  • Control frequency: 1kHz (force-position hybrid)
  • Quick-change end effector interface

3.2 Force Control: From “Factory” to “Operating Room”

Three key modifications by UCSD:

  1. Impedance control: High damping (20 Ns/m) for stability, low stiffness (500 N/m) for compliance
  2. Force limiting: Max 3N to protect tissue
  3. Tissue stiffness estimation: Real-time adaptation based on force history
// Surgical force controller core
type ImpedanceController struct {
	Damping   float64 // 20 Ns/m (surgical: high damping)
	Stiffness float64 // 500 N/m (surgical: low stiffness for compliance)
	MaxForce  float64 // 3.0 N (tissue protection)
}

func (ic *ImpedanceController) ComputeForce(desired, actual [3]float64) [3]float64 {
	var force [3]float64
	for i := 0; i < 3; i++ {
		posErr := desired[i] - actual[i]
		force[i] = ic.Damping*0 + ic.Stiffness*posErr // simplified
		if force[i] > ic.MaxForce { force[i] = ic.MaxForce }
	}
	return force
}

4. Dual-Robot Cooperation: 32 Minutes, No Human Intervention

The most impressive result: two G1 robots performing the complete surgery without human intervention, 42% faster than single-robot operation.

Coordination Protocol:

  • Task allocation: Lead robot handles critical operations (cutting, clipping); assistant handles support (exposure, suction)
  • Collision avoidance: Real-time trajectory sharing; assistant automatically yields to lead
  • Emergency interlock: Either robot triggers stop → both stop simultaneously
  • State synchronization: Shared coordinate system; real-time force data exchange

5. Implications

5.1 Breaking Da Vinci’s Monopoly

With over 8,000 installations worldwide, Da Vinci surgical systems have long dominated robotic surgery. A single Da Vinci costs $2-4M plus $150K/year maintenance. One G1 costs ~$40K. The same investment in Da Vinci could deploy 100 G1 robots.

5.2 The “MD Moment” for Humanoid Robots

This is the first empirical proof that a general-purpose platform can exceed a specialized system in a professional domain. If a 27kg G1 can perform cholecystectomy, other humanoids (H1, Figure 02, Optimus) with similar force control and vision systems can theoretically do the same.

5.3 Future of Surgery

Remote surgery, autonomous surgery, cooperative surgery — these are no longer science fiction. A 32-minute autonomous dual-robot cholecystectomy is just one step away from truly autonomous surgical robots.


6. Conclusion

The Unitree G1’s Nature publication on live porcine cholecystectomy is a landmark event in humanoid robot history. It proves:

  1. Low-cost general platforms can challenge specialized systems: A $40K, 27kg G1 performed tasks only a $2-4M, 500kg Da Vinci could do
  2. Force control precision is key: Industrial-grade force control can be upgraded to surgical-grade through algorithm modification
  3. Dual-robot cooperation is the future: 32-minute autonomous surgery demonstrates the potential of humanoid “swarm intelligence”
  4. Healthcare is the largest blue ocean for humanoids: From surgery to nursing, rehabilitation to companionship, the medical market dwarfs industrial and domestic service