Google Gemini AI Powers Boston Dynamics Spot for Enhanced Autonomy

Boston Dynamics announced a partnership with Google Cloud and Google DeepMind on April 14, 2026, to integrate the Gemini AI system into its Spot robot. The integration, which utilizes Gemini and Gemini Robotics-ER 1.6 within the Orbit AIVI-Learning platform, is designed to enhance the robot’s reasoning and adaptability in industrial environments.

By implementing Gemini as a high-level reasoning engine, Spot can now execute complex tasks by calling upon various tools and models. This allows the quadruped robot to perform higher-order visual analysis and task planning, moving beyond simple instruction-following to a more intuitive AI experience.

Technical Capabilities of Gemini Robotics-ER 1.6

Alongside the partnership announcement, Google DeepMind released Gemini Robotics-ER 1.6 on April 14, 2026. This reasoning-first model is specifically engineered for embodied reasoning, which is the process of bridging digital intelligence with physical action to help robots reason about the physical world.

The model introduces several technical improvements over previous versions, including Gemini Robotics-ER 1.5 and Gemini 3.0 Flash. These enhancements focus on spatial and physical reasoning, specifically in areas such as pointing, counting, and success detection.

Gemini Robotics-ER 1.6 is capable of natively calling external tools to complete tasks, including:

• Google Search for retrieving information
• Vision-language-action models (VLAs)
• Third-party user-defined functions

Industrial Application and Instrument Reading

The collaboration between Boston Dynamics and Google DeepMind led to the discovery and development of a new capability known as instrument reading. This allows robots to interpret and read complex sight glasses and gauges, a critical requirement for monitoring industrial equipment.

The integration into Orbit AIVI-Learning is expected to improve Spot’s performance across several critical industrial functions, including:

• Asset monitoring
• Materials management
• Safety and security checks

Boston Dynamics intends for these AI models to provide higher accuracy and a better baseline of performance that can improve continuously without the need for manual software updates.

Developer Access and Broader Context

Google DeepMind has made Gemini Robotics-ER 1.6 available to developers as of April 14, 2026, through the Gemini API and Google AI Studio. To facilitate adoption, the company released a developer Colab featuring examples of how to prompt the model for embodied reasoning tasks and configure its settings.

This partnership expands upon an existing relationship between the two companies. In addition to the work on Spot, Google DeepMind has also been collaborating with Boston Dynamics to increase the intelligence of the Atlas humanoid robot.

For robots to be truly helpful in our daily lives and industries, they must do more than follow instructions, they must reason about the physical world.

Google DeepMind