From Thinking AI to Acting AI: The Rise of Physical AI

Session Overview

Generative AI has made remarkable progress in understanding and reasoning across language and images. Physical AI takes the next step by extending that intelligence into action in the real world.

With the emergence of vision-language-action models and robot foundation models that translate visual and linguistic information into physical movement, robots are evolving beyond machines that repeat predefined tasks. They are becoming general-purpose systems capable of responding to diverse environments, situations, and instructions.

However, acting reliably in the physical world requires more than simply recognizing a scene. AI systems must understand movement, remember past situations, and make use of physical signals such as touch and force.

This session examines the core technologies behind Physical AI, the changing role of data and learning methods, and the key challenges that must be solved for robots to operate safely and consistently beyond the laboratory—in real industries and everyday life.

Key Takeaways

• How AI is evolving from understanding and reasoning to acting in the physical world
• The role of vision-language-action models and robot foundation models in general-purpose robotics
• Why movement understanding, memory, touch, and force sensing are essential for Physical AI
• The data, learning, reliability, and safety challenges that must be solved for real-world deployment