A Google DeepMind main speaker explains how Gemini Robotics 2 aims to help robots handle varied physical tasks, including coordinated movement, dexterous manipulation, and teamwork between robots.
The speaker describes everyday interaction with the physical world as intuitive for people but extremely difficult for robots. Gemini Robotics 2 is designed as a general-purpose model, with the aim of letting one robot perform many different tasks.
The Gemini Robotics system can control a humanoid's body, a Sharpa hand, and Duo's grippers. Whole-body control requires coordinating many decisions across the robot, even for movements that feel simple to people.
A second focus is fine hand control for tasks such as screwing in a light bulb. The trash-bag task illustrates the challenge: a hand with 22 joints must be controlled without relying on the intuitive coordination people take for granted.
Gemini Robotics 2 supports multiple robots working on a task at the same time. In the tool-kit example, Apollo asks Duo to gather tools, close the kit, and return it to a bin; each robot runs its own copy of the system and coordinates through reasoning.
The speaker emphasizes reacting when the scene changes as a key part of robot intelligence and generality. They describe AI as a missing piece for robotics and say the field is advancing quickly.
Briefs summarise talks and interviews; timestamps open the video at that moment. Claims in a talk are the speaker's own.
TAKASU Masakazu, host of Nico-Tech Shenzhen, reports from ROSCon China 2026 in Shenzhen on the robotics demos, developer tools, and embodied AI themes across the event.
A Sharpa presenter introduces the company’s dexterous robot hands, CraftNet AI model, data strategy, and real-world deployments.
Scott Walter, a robotics engineer and advisor, joins Over The Horizon hosts and guests to discuss Sharpa’s Dairy Queen deployment, Mimic’s dexterous hand, and Sunday Robotics’ home laundry results.
Shiao-Ju, who works on tactile sensing and related algorithms at Sharpa, presents at GTC 2026 on using tactile AI for high-degree-of-freedom dexterous manipulation.
