Brains, Bodies and Benchmarks: Closing the Robotics Gap in 2026

​Physical AI is at an inflection point

​​Model, data and hardware breakthroughs have converged for AI to perceive, reason and act in the physical world. Join industry leaders from top robotics companies for a panel discussion on what comes next.

​​Co-hosted by Encord, Weights & Biases, and Anyscale - the end-to-end MLOps stack for scaling embodied intelligence.

​What We'll Cover:

• ​​Data, training and inference challenges teams face scaling robotics systems

• ​​How recent breakthroughs (e.g. VLM/VLA, UMI, World Models, etc) are reshaping robotics design, training and deployment

• ​​Validation and benchmarking approaches to find the best models 

• ​How new developments in robotics are impacting AI teams around the world

​Panelists: 

• ​Chris Paxton (AI Innovation Lead at Agility Robotics) 

• ​Jason Ma (Co-Founder at Dyna Robotics)

• ​Daniel Ho (Director of Evaluations at 1X)

• ​Adrian Li-Bell (Member of Technical Staff at Physical Intelligence)

​Moderated by:

• ​Chris Paulson (Senior DevRel Manager, Robotics at NVIDIA)

​Casual networking with refreshments before and after the panel discussion.

​Hosts

​​​​​Encord helps leading Physical AI teams accelerate data iteration cycles. With a unified platform for data curation, annotation and model evaluation across 2D and 3D data, teams can train better models with high quality, representative data at scale.

​​Weights & Biases is the AI developer platform. Build better models faster, fine-tune LLMs, develop GenAI applications with confidence, all in one system of record developers are excited to use.

​​Anyscale enables Python developers to build and run all their AI—from data prep to training and inference—at any scale. Anyscale is trusted by leading AI teams at Canva, TripAdvisor, Physical Intelligence, Coinbase and more.