Innovation Brief · IB-2025-07c · July 2025
Physical AI
How AI and LLMs are supercharging robotics
Generative AI is producing a new class of intelligence that leaves the screen — robots aware of their surroundings and able to act on instructions given in natural language.
- Authors
Eugene ChevyrovInnovation Engineering
Nick DiazInnovation Engineering
Ian KriegerAPAC Innovation Officer
- Length
- 16 pages
- Themes
- Physical AI, The Future of AI, Smart Spaces
Summary
- Generative AI is accelerating robotics development, leading to a new class of AI — physical AI — that is powering far more capable and versatile robots.
- AI is driving the emergence of robots both aware of their surroundings and capable of taking sequences of actions based on high-level instructions in natural language.
- In the enterprise, one of the most promising applications of physical AI will be combining it with digital AI, enabling workflows to extend into the physical world.
- As physical AI advances, a platform approach will be vital to coordinating workflows across humans, digital AI agents and physical AI agents.
- Physical AI will unlock new business possibilities by powering robots capable of handling complex tasks as a seamless part of operations.
- Investing in physical AI and a platform approach today puts businesses at the forefront of the next frontier in artificial intelligence.
From the brief
Robots have long been long on promise but limited to applications such as factory lines, kitchens, warehouses, hospitality venues and restaurants — largely confined to single, repetitive tasks in controlled settings.
All of that is about to change. Physical AI allows robots to interpret their surroundings and take actions based on that understanding, overcoming the historic barriers that kept robots unaware and inflexible.
Read the full report