Physical AI Advances Require New Foundation Models
At CES 2026 and in interviews with The New Stack, industry leaders from Nvidia, Boston Dynamics, and Bedrock Robotics outline advances in physical AI. They say new foundation models — including large behavior models, vision-language-action models, open-world models, and generative policies — are required for perception, world modeling, and tactile control. Enterprises must integrate hardware, edge-capable models, and lifecycle tooling to scale adoption.
Key Points
- 1Identify adoption of new foundation model classes (LBMs, VLAs, open-world models) for physical AI
- 2Enable perception, world modeling, and tactile control under unpredictable real-world conditions
- 3Advise enterprises to build integrated hardware-software stacks, edge-capable models, and lifecycle tooling
Scoring Rationale
Industry-backed overview of emergent physical-AI foundation models; offers practical guidance but lacks original empirical evaluation.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
