Physical AI: the new era of robotics
Google DeepMind and Boston Dynamics are partnering to bring frontier AI breakthroughs into physical robotics
“you have this vision language and action model and you could like talk to the robot and you can ask it to do things and you get all this like understanding for free in robotics”
At Google I/O 2026, Google DeepMind and Boston Dynamics leaders described how large multimodal models are being extended with 'action tokens' to create vision-language-action (VLA) models that give robots emergent real-world understanding. The partnership between DeepMind and Boston Dynamics signals a major convergence of frontier AI and physical robotics hardware. The speakers framed 2026 as the inflection year where general AI breakthroughs translate meaningfully into robot capabilities.