Architecting Physical AI SoCs with Standards‑Based IP for Real‑World Intelligence

Featured Speakers:

• Hezi Saar, Executive Director, Synopsys

As AI systems move beyond the cloud and into the physical world, SoC architectures must evolve to support real‑time perception, decision‑making, and interaction with their environments. 

This session explores how:

• edge and Physical AI are reshaping autonomous systems. 
• multi-modal sensing and high-performance compute enable reinforcement learning in the real world. 
• standards-based IP provides a scalable foundation for Physical AI SoCs. 
• JEDEC UFS/LPDDR, MIPI CSI/DSI/I3C, PCIe, and security IP fit into modern SoC architectures. 
• these interfaces deliver the bandwidth and latency needed for sensor fusion and AI inference. 
• determinism and reliability support real-time control under physical constraints. 
• security and safety-aware design principles strengthen deployable Physical AI platforms.