Boston Dynamics Robot Dog Gets Upgraded With Gemini AI for Near-Autonomous Work

Boston Dynamics has enhanced its quadruped robot Spot, by integrating advanced AI from Google DeepMind, aiming to improve its ability to operate autonomously in industrial environments.

AI Integration

Spot now runs Gemini Robotics-ER 1.6, an embodied reasoning model designed to help robots better understand and interact with the physical world. The upgrade allows the robot to perform inspection tasks with greater independence and accuracy.

Boston Dynamics is among the few companies deploying legged robots at scale, with several thousand units already in use across industries.

Focus on Industrial Inspections

The integration targets inspection work, one of the most practical applications for legged robots. Spot is used to navigate industrial sites and monitor conditions, including identifying potential hazards.

With the new AI system, the robot can detect dangerous debris or spills, read complex gauges, and interpret sight glass indicators. These tasks previously required more human oversight.

Improved Reasoning and Decision-Making

The addition of Gemini Robotics-ER 1.6 enables Spot to decide when to use additional AI tools. It can call on vision-language-action models when it encounters situations that require deeper analysis.

This layered approach allows the robot to better assess its surroundings and respond to unexpected conditions.

According to Boston Dynamics, the upgrade improves Spot’s ability to observe, interpret, and act without direct human input. The company says these advancements support more reliable task execution in real-world environments.