New NVIDIA Jetson Thor Computers Expand Mainstream Robotics and Edge AI
NVIDIA just introduced new Jetson Thor computers, built to put physical AI into mainstream robots.
Transcript
NVIDIA just introduced new Jetson Thor computers, built to put physical AI into mainstream robots.
Jetson Thor is the runtime brain of NVIDIA's robot stack. It works with DGX for training and Omniverse for simulation.
The original chip delivers up to seven point five times the performance of the last generation, running large models fully on device.
Launched last August. A software update in October brought three point three times faster LLM speed. Now, new models push it wider.
Mainstream Jetson Thor means AI agents can run at the edge, inside more robots, cameras, and machines everywhere.
We make Inboxsmith, the AI receptionist that never misses a business call. Please like and subscribe for more news.
Sources
Every claim in this video comes from the top ranking coverage of this topic. The claims and where each one came from:
- Jetson Thor is part of NVIDIA's three-computer solution: DGX for training, Omniverse for simulation, Jetson Thor for runtime robotics.(Getting Started with the NVIDIA Jetson AGX Thor Developer Kit for Physical AI)
- Up to 7.5x the performance of Jetson AGX Orin, runs modern AI models including LLMs and VLMs on device.(Getting Started with the NVIDIA Jetson AGX Thor Developer Kit for Physical AI)
- October 2025 software update delivered 3.3x LLM throughput gains via flash infer and XFormers integration.(Jetson Thor LLM Performance Gains - Up to 3.3x Faster!)
- Jetson AGX Thor developer kit launched August 25, 2025.(Getting Started with the NVIDIA Jetson AGX Thor Developer Kit for Physical AI)