NVIDIA Reveals Its Most Inexpensive Supercomputer for Generative AI.
NVIDIA Jetson Orin Nano
- With a generative AI performance boost of up to 1.7x, the Jetson Orin Nano Super supports well-liked models for educators, developers, and hobbyists.
- With a software update, NVIDIA is revealing a new small generative AI supercomputer that offers better performance at a reduced cost.
This version improves generative AI inference performance by 1.7x, performance by 70% to 67 INT8 TOPS, and memory bandwidth by 50% to 102GB/s. It is now available.
The Jetson Orin Nano Super is the perfect tool for developing retrieval-augmented generation-based LLM chatbots, developing a visual AI agent, or deploying AI-based robotics.
The Present That Never Stops Giving
For users of the Jetson Orin Nano Developer Kit, the software upgrades for the new Jetson Orin Nano Super will also improve generative AI performance.
Those who want to learn more about computer vision, robotics, or generative AI can choose the Jetson Orin Nano Super. AI offers an approachable framework for turning concepts into reality as it transitions from task-specific models to foundation models.
Strong Performance Using Generative AI Super
All well-known generative AI models and transformer-based computer vision benefit from the Jetson Orin Nano Super’s improved performance.

A reference carrier board and an 8GB Jetson Orin Nano system-on-module (SoM) make up the development kit, which offers the perfect foundation for developing edge AI applications.
High-performance inference and numerous concurrent AI application pipelines are made possible by the SoM’s 6-core Arm CPU and NVIDIA Ampere architecture GPU with tensor cores. It offers better resolution and frame rates than earlier iterations and supports up to four cameras.
Wide-ranging Community and Software Ecosystem for Generative AI
The field of generative AI is rapidly developing. For those state-of-the-art models from the open-source community, the NVIDIA Jetson AI lab delivers instant assistance along with user-friendly lessons. Additionally, developers can get inspiration from developer-made projects and a wealth of help from the larger Jetson community.
NVIDIA AI software, such as NVIDIA Holoscan for sensor processing, NVIDIA Metropolis for visual AI, and NVIDIA Isaac for robotics, is used by Jetson. With NVIDIA Omniverse Replicator for creating synthetic data and NVIDIA TAO Toolkit for optimizing pretrained AI models from the NGC catalogue, development time may be decreased.
Additional AI and system software, developer tools, and bespoke software development are provided by Jetson ecosystem partners. Along with carrier boards and product solution design, they can also assist with cameras and other sensors.
NVIDIA Jetson Orin Nano Super Developer Kit
The cheapest supercomputer for generative AI.
Crucial Performance
- 67 TOPS for AI Performance
- 102 GB/s of memory bandwidth
- CPU Speed: 1.7 GHz from 40 TOPS up. from 68 GB/s. over 1.5GHz.
Workloads
Open Up New Edge Physical AI Applications
Edge of Generative AI
NVIDIA Jetson Orin: Providing effective and powerful generative AI.
For the newest generative AI applications, NVIDIA Jetson Orin Nano Super offers outstanding AI computation, quick memory bandwidth, and extensive software stacks with great energy efficiency. For any transformer-based generative AI model, it facilitates quick inference.
Robotics
Accelerate robots powered by AI in a variety of sectors.
Use AI to speed up the development, simulation, and deployment of robots, allowing essential capabilities like vision, grasping, and mobility. Construct robots for a variety of sectors, including as industry, retail, healthcare, logistics, agriculture, and more.
AI Vision
Use vision AI to turn data from trillions of IoT devices into insightful knowledge.
NVIDIA Metropolis combines AI and visual data to increase productivity and security across a range of sectors. It handles sensor data for uses like as healthcare, smart city traffic, inventory control, frictionless retail, and factory inspections. Create, implement, and scale AI and IoT applications with Metropolis from the edge to the cloud.
Advantages
Developing Generative AI on the Boundaries
Improved Performance for Modern Generative AI Models
NVIDIA’s latest software upgrade boosts Jetson Orin Nano’s AI performance from 40 to 67 AI TOPs and memory bandwidth from 68 to 102 GB/s, making it easier for developers and fans to run generative AI models.
All Jetson Orin Nano Developer Kits Have a Super Increase
Through a software upgrade, any currently available Jetson Orin Nano Developer kits in the market may achieve this “Super” performance boost.
Performance
Revolutionary Results for Generative AI Models
On several well-known Large Language Models (LLMs), Vision Language Models (VLMs), and Vision Transformers, the NVIDIA Jetson Orin Nano Super offers notable performance gains over its predecessor.
Jetson Orin Nano Super Developer Kit price
Strong and Reasonably Priced
The most potent and reasonably priced embedded computer for generative AI is the Jetson Orin Nano Super Developer Kit. It is now available for purchase from NVIDIA authorized distributors globally for $249 USD.
The palm-sized NVIDIA Jetson Orin Nano Super Developer Kit improves generative AI performance and capabilities for enterprise AI developers, enthusiasts, and students. The price reduced from $499 to $249.
Jetson Orin Nano Super Developer Kit Getting Started
Start Using the Jetson Orin Nano Super Developer Kit
With the Jetson Orin Nano Super Developer Kit, Begin Your Adventure
One of the global partners is offering the new Jetson Orin Nano Super Developer Kit for purchase.
Jetson Orin Nano Specs
Jetson Orin Nano Super Developer Kit Specs

AI Performance | 67 INT8 TOPS |
GPU | NVIDIA Ampere architecture with 1024 CUDA cores and 32 tensor cores |
CPU | 6-core Arm Cortex-A78AE v8.2 64-bit CPU 1.5MB L2 + 4MB L3 |
Memory | 8GB 128-bit LPDDR5 102 GB/s |
Storage | Supports SD card slot and external NVMe |
Power | 7W–25W |
Category | Specifications |
---|---|
Camera | 2x MIPI CSI-2 22-pin camera connectors |
PCIe Slots | – PCIe M.2 Key M slot with x4 PCIe Gen3 |
– M.2 Key M slot with x2 PCIe Gen3 | |
– M.2 Key E slot | |
USB Ports | – USB Type-A connector: 4x USB 3.2 Gen2 |
– USB Type-C connector for UFP | |
Networking | 1x GbE connector |
Display | 1x DP 1.2 (+MST) connector |
Other I/O | – 40-pin expansion header (UART, SPI, I2S, I2C, GPIO) |
– 12-pin button header | |
– 4-pin fan header | |
Power | DC power jack |
Mechanical | – Dimensions: 103mm x 90.5mm x 34.77mm |
– Height includes feet, carrier board, module, and thermal solution |