Canonical
on 27 August 2025
Canonical announced that official support for the NVIDIA Jetson Thor family is coming to Ubuntu, continuing a strategic collaboration with NVIDIA to accelerate AI innovation at the edge. The official support from Canonical will offer optimized Ubuntu images and enterprise-grade stability and security. This commitment to long-term support and security updates ensures enterprise-grade stability and reliability to the combination of Ubuntu and NVIDIA Jetson system-on-modules.
NVIDIA Jetson Thor: A powerful edge platform for physical AI and robotics
NVIDIA has just announced the availability of the Jetson AGX Thor Developer Kit along with the Jetson Thor™ series modules. The NVIDIA Jetson AGX Thor Developer Kit is a powerful platform for humanoid robotics and physical AI applications, delivering unmatched performance and scalability in a compact, power-efficient form factor. Built on advanced NVIDIA Blackwell and 128 GB of memory, NVIDIA Jetson Thor series modules deliver up to 2070 FP4 TFLOPS of AI compute effortlessly running the latest generative AI models to enable highly sophisticated robotic systems – all in a compact form factor. Compared with its predecessor, NVIDIA Jetson Orin™, Jetson Thor delivers up to 7.5x higher AI compute and 3.5x greater energy efficiency.
The NVIDIA Jetson Thor developer kit is a complete system for software development and prototyping, featuring a module on a carrier board with standard I/O. The Jetson Thor module is the core computing component designed for integration into a final product’s custom hardware. The developer kit is for creating and testing software, while the module is for deploying that software in a production device.
Committing to secure & reliable AI deployment at the edge
NVIDIA JetPack is a comprehensive software development kit (SDK) that provides a full development environment for building and deploying AI applications on the NVIDIA Jetson platform. JetPack 7 uses the Linux 6.8 kernel with real-time support, which is the same as in Ubuntu 24.04 LTS. Canonical’s official Ubuntu image for NVIDIA Jetson Thor will enable developers and enterprises to create breakthrough AI solutions, such as vision AI agents, Generative AI edge applications, medical devices and autonomous robots. The work between Canonical and NVIDIA will enable sustainable deployment of Generative AI applications at the edge – for the entire Jetson Thor series of modules.
Key benefits of this collaboration include:
- Enterprise-ready security: Benefit from Ubuntu’s robust security features and long-term support, which enables compliance with regulations such as the Cyber Resilience Act.
- Reliable stability on certified hardware: Canonical’s QA team performs an extensive set of over 500 OS compatibility-focused hardware tests to ensure that every aspect of the system is checked and verified for the best Ubuntu experience.
- Seamless development to deployment: Use the same architecture in your workstation and the cloud for a clear path to mass edge AI application distribution.
What does it mean for application developers?
The NVIDIA Jetson Thor modules are optimized for physical AI and robotics. Developers will be able to run the most popular GenAI models locally, such as Llama3.1, qwen3. gemma3, and deepseek-r1. Application developers can use their familiar inference engines like Ollama, vLLM, and Huggingface.
What makes Jetson Thor modules transformative is their support for GenAI applications. Its features include:
- Real-time kernel: A real-time kernel ensures predictable, low-latency task execution. This is critical for robotics applications, which need deterministic timing to avoid jitter.
- Multi-instance GPU (MIG): MIG partitions a single GPU into multiple secure, isolated instances. Each instance has dedicated hardware resources, allowing developers to run multiple AI models simultaneously with guaranteed performance.
- Support Any Generative AI model and Framework: JetPack7 provides a powerful compute stack to support any model AI models including large language model, vision language model or vision language action model such as NVIDIA GR00T N.
- NVIDIA Holoscan: This platform enables low-latency data streaming from sensors directly to the GPU. It allows developers to perform real-time sensor fusion and perception without CPU bottlenecks.
A shared vision for the AI era
This Jetson Thor release is the latest in a long series of Canonical initiatives integrating NVIDIA technology, and one of many releases under Canonical’s Silicon Partner Program, which releases optimized images of Ubuntu for certified hardware from the world’s biggest manufacturers.
Canonical’s work with NVIDIA technologies aligns perfectly with our mission to make AI accessible for developers to build and reliable for businesses at scale. As a member of the NVIDIA Partner Network (NPN), Canonical offers official support for a wide range of NVIDIA products and solutions. Whether you’re an enterprise delivering commercial solutions or an enthusiast looking to get hands-on with AI, Ubuntu on NVIDIA Jetson Thor provides the ideal platform for building the next million-device AI application.
Getting started
Jetson Thor image availability will be announced soon.
To get started with Ubuntu on NVIDIA Jetson, visit our download page.
These images are optimized for both the developer kit and the modules, ensuring you can build the most powerful edge AI applications, and scale their deployment successfully. Using these images is fast and easy: simply select the OS image to match your hardware, flash it onto a USB or NVMe disk, and load it onto your board. For Ubuntu Core, you will be able to download pre-built images for exploration and experimentation.
About Canonical
Canonical, the publisher of Ubuntu, provides open source security, support, and services. Our portfolio covers critical systems, from the smallest devices to the largest clouds, from the kernel to containers, from databases to AI. With customers that include top tech brands, emerging startups, governments and home users, Canonical delivers trusted open source for everyone.
Learn more at https://canonical.com/
Learn more
Find out more about Canonical’s collaboration with NVIDIA.
