FrontierJuly 31, 2026via The Decoder
Google Deepmind unveils Gemini Robotics 2 to power robots of all shapes from tabletop arms to humanoids
Why it matters
A new frontier model class: multimodal AI that reasons about and controls physical systems at scale. This is the capability layer that turns robotics from siloed research into a generalizable platform problem.
Key signals
- Gemini Robotics 2 is a vision-language-action model
- Designed to control robots across form factors: tabletop arms, mobile bases, full-body humanoids
- Includes higher-level reasoning layer for robotics task decomposition
- Published July 31, 2026
- From Google Deepmind (lab-race relevant: Gemini is Google's frontier model line)
The hook
Google Deepmind ships Gemini Robotics 2—a vision-language-action model that unifies control across tabletop arms to full humanoids.
Google Deepmind's Gemini Robotics 2 is its most advanced vision-language-action model yet, built to control everything from tabletop robots to full-body humanoids. Gemini Robotics ER 2 adds a higher-level reasoning layer for robotics tasks.