Tesla AI Chief Elluswamy to Keynote at CVPR 2026 in Denver

AUSTIN, Texas — Tesla's head of Autopilot and AI, Ashok Elluswamy, is set to take the stage at the 2026 Conference on Computer Vision and Pattern Recognition (CVPR), the most prestigious academic gathering in computer vision, running June 3–7 in Denver, Colorado.

According to the confirmed CVPR 2026 schedule, Elluswamy is presenting at three separate workshops on June 3 at the Colorado Convention Center. The sessions span the Workshop on Autonomous Driving (WAD), the AUTOPILOT Workshop, and the inaugural Workshop on Deployment of Foundation Models for Embodied AI.

A Title Worth Noting

The title of Elluswamy's talk at that third workshop — "Building Foundational Models for Robotics at Tesla" — carries specific weight. It frames Tesla's autonomous driving work not as a standalone vehicle product, but as part of a broader robotics and embodied AI platform that also encompasses the Optimus humanoid robot program. It signals that Tesla's AI architecture is increasingly unified across domains rather than siloed by application.

A Side-by-Side with XPENG

The lineup at CVPR puts Elluswamy back-to-back with XPENG's head of General Intelligence, creating what may be the most revealing public comparison between two of the world's most widely deployed vision-based end-to-end autonomous driving systems. Both companies build camera-only, end-to-end neural networks trained on fleet data — but this will be one of the first times their approaches have been formally presented in the same academic forum.

The AUTOPILOT Workshop specifically focuses on open-world perception and integrated language models for on-road tasks, mapping directly to the direction Tesla's Full Self-Driving stack has been heading since v14.

What It Means for Tesla's AI Strategy

For followers of Tesla's AI development, CVPR appearances by senior leadership are relatively rare. When they do happen, they tend to reveal meaningful technical detail about the underlying architecture. Elluswamy's prior CVPR presentations — including talks on Occupancy Networks and end-to-end neural planners — preceded real-world FSD improvements by months.

The academic community's engagement with Tesla's methodology also tends to accelerate publication and independent benchmarking, which has historically benefited Tesla by surfacing improvements the team can incorporate into future training runs.

With Tesla's unified AI model now reportedly shared between FSD, Optimus, and the Cybercab robotaxi platform, whatever Elluswamy presents in Denver will offer the clearest public window yet into how the company is thinking about general-purpose machine intelligence — and what comes next.