Physical AI Symposium
👉🏼 The Physical AI Symposium 2026 brings together researchers, engineers, and entrepreneurs pioneering the intersection of AI, spatial computing, and embodied interaction.
🫡 Powered by Phygtl, Inc. in collaboration with leading universities and industry partners, the Assembly serves as a forum for innovation and applied research around the emerging paradigm of
+ Large World Models (LWM)
+ AI systems capable of understanding real-world dynamics and human context.
🎯 Through keynotes, collaborative sessions, and immersive showcases, participants will share breakthroughs in 3D vision, multi-agent reasoning, and human-AI co-creation.
PHYSICAL AI SYMPOSIUM 2026 BY PHYGTL
AGENDA
From World Models to Living Experience
Palo Alto, California
Thursday, April 23, 2026
9:15 AM – 2:30 PM (PDT)
Invite-Only · Closed-Door In-Person Assembly
Global Academic Audience (Remote Access)
ABOUT THE SYMPOSIUM
The Physical AI Symposium 2026 convenes a highly curated group of researchers, system builders, and founders shaping the next generation of AI systems that understand, model, and interact with the physical world.
Hosted in Palo Alto, the symposium focuses on the emerging paradigm of:
Large World Models (LWM) — persistent, spatial, dynamic representations of reality
Human-centered perception and interaction
Embodied and agentic AI systems
The transition toward Large Experience Models (LEM)
09:15 - 09:40 : OPENING KEYNOTE — FOUNDATIONS
When AI Meets the Physical World
Why perception — not language — is the real bottleneck
Category: Vision · Geometry · Foundations
09:40 – 10:05 : CORE REPRESENTATION SHIFT
Gaussian Splatting, Neural Fields & the New 3D
Paradigm
How machines represent space now
Category: 3D / 4D Perception · Scene Representation
10:05 – 10:30 : FROM PAPERS TO REAL TIME
From Static Scenes to Living Worlds
Time, scale, and real-time constraints
Category: 4D Perception · Systems · Infrastructure
10:30 – 10:50 — Break
Visual demos: neural scenes, volumetric capture, avatars.
10:50 – 11:15 : HUMANS IN THE LOOP
Humans Are Not Objects
Perceiving people, motion, and intent
Category: Human-Centered Perception · Interaction
11:15 – 11:40 : EMBODIED INTERACTION
Perception Is Not Enough
From seeing to acting
Category: Embodied AI · Agents · Perception-Action Loops
11:40 – 12:05 : GENERATIVE WORLDS
Generative Worlds Without Hallucination
Constraining creativity with physics
Category: Generative AI · World Modeling
12:05 – 12:45 — Closed Lunch
In-room participants only.
12:45 – 13:10 : THE EXPERIENCE LAYER
Toward the Large Experience Model (LEM)
From world understanding to world creation
Category: Experience · Interaction · Emergence
13:10 – 14:00 : CLOSING FIRESIDE
What Are We Still Getting Wrong About Physical AI?
Format:
Moderated fireside · 2 speakers max · No slides
14:00 – 14:30