Humanoid Robot Demo to Debut at Austin Keynote
Nvidia's Project GR00T team will show autonomous humanoid hardware at CoRL 2026, a research preview rather than a shipped product.
Updated on Nov. 7, 2026 in Robotics

Jim Fan, who co-leads Nvidia's Project GR00T, has confirmed that the teaser for his keynote at the Conference on Robot Learning (CoRL) 2026 shows real humanoid hardware running autonomously at normal speed. The keynote is scheduled for November 9, 2026, in Austin, Texas. This is a forecast of a scheduled talk, not a shipped product.
Why it matters
Robots lack an internet-sized library of physical movement to learn from, and collecting real robot data is slow and expensive, which is the core obstacle Project GR00T targets. The keynote is a checkpoint on whether Nvidia's approach of expanding limited data with simulation and generative world models can produce autonomous humanoid behavior.
GR00T N1 is a 2B-parameter model, and the project's method expands limited teleoperation data by over 1,000x through simulation and generative world models. The teaser's autonomous, real-time footage has not yet been independently measured against any benchmark, and the keynote is expected to explain the training pipeline.
The players
Jim Fan
Co-leads Nvidia's Project GR00T, the company's humanoid robot foundation-model effort.
Nvidia
GPU and AI platform company whose Isaac Lab and Cosmos tools underpin its robot-learning pipeline and GR00T models.
Jensen Huang
Nvidia CEO who unveiled Project GR00T at GTC 2024.
Conference on Robot Learning (CoRL)
Annual research conference on robot learning hosting the 2026 keynote in Austin, Texas.
The details
Project GR00T is Nvidia's effort to build foundation models for humanoid robots, meaning general-purpose models meant to be adapted across many tasks rather than programmed for one. Because real robot demonstrations are scarce, the team uses teleoperation, in which a human remotely controls a robot to record examples, and then multiplies that small dataset using simulated environments and generative world models, which are AI systems that synthesize plausible scenes and motion. Platforms such as Isaac Lab and Cosmos support this pipeline, which the keynote is expected to describe.
Timeline
March 2025: Nvidia introduced the open-weight GR00T N1 model.
2025: Follow-up versions N1.5 and N1.6 were expected to roll out.
November 9, 2026: Jim Fan's CoRL 2026 keynote in Austin, Texas.
The Tech Race
Nvidia's GR00T line follows its 2024 unveiling and its March 2025 open-weight N1 release, positioning the company as a platform supplier for humanoid learning. The CoRL 2026 keynote gives the research community a first look at whether that platform strategy yields autonomous hardware at normal speed.
For now this is a preview of research rather than a product; no availability or pricing has been announced for GR00T-based robots. Readers in Austin can watch the November 9 keynote, while practical access to humanoid platforms built on this work remains unscheduled.
The takeaway
Watch the November 9, 2026 CoRL keynote in Austin for the task range and training-pipeline details that will show whether autonomous humanoid behavior holds up beyond the teaser. Track whether the follow-up N1.5 and N1.6 releases, expected through 2025, appear as the next measurable milestones.
What happens next
Jim Fan's CoRL 2026 keynote is scheduled for November 9, 2026, in Austin, Texas, where the range of tasks and the training pipeline are expected to be presented.
Further reading
For the broader research and industry context, explore Robotics coverage in Austin.




