NVIDIA Jetson Orin Nano Super Developer Kit

Edge-AI dev kit delivering up to 67 TOPS with a 6-core Arm CPU and Ampere GPU for generative AI and robotics.

₹37,200 + 18% GST

SKU: BOTBIT-DEV-15 · Best for: Running local generative AI and small language models (LLMs/SLMs) at the edge without cloud

Qty

Overview

About NVIDIA Jetson Orin Nano Super Developer Kit.

The NVIDIA Jetson Orin Nano Super Developer Kit is a compact edge-AI computer designed for building and running generative AI, computer vision and robotics applications right on the device. With the 'Super' software update, the same hardware unlocks a boosted power mode that delivers up to 67 INT8 TOPS of AI performance, making it one of the most capable entry-level AI dev kits you can put on a robot or an autonomous camera.

It pairs a 6-core Arm Cortex-A78AE CPU with a 1024-core NVIDIA Ampere GPU (32 Tensor Cores) and 8GB of LPDDR5 memory, and it runs the full NVIDIA JetPack SDK with CUDA, cuDNN and TensorRT. That means you can run local LLMs, vision transformers, YOLO object detection and ROS 2 robotics stacks without depending on the cloud.

For Indian makers, students, startups and robotics labs, it has become the go-to platform for serious edge-AI work because it is dramatically faster than a Raspberry Pi for neural networks while staying far cheaper than a full GPU workstation. The kit includes the module pre-mounted on a reference carrier board, so you can boot straight into Ubuntu-based Linux.

Key features

Why it stands out.

Up to 67 TOPS (INT8) AI performance with the free Jetson 'Super' JetPack update, roughly 1.7x the original Orin Nano throughput

1024-core NVIDIA Ampere architecture GPU with 32 Tensor Cores for accelerated deep-learning inference

6-core Arm Cortex-A78AE 64-bit CPU running at up to 1.7 GHz

8GB 128-bit LPDDR5 memory delivering 102 GB/s bandwidth for large models

Configurable 7W to 25W power profiles so you can trade performance for battery life on mobile robots

Rich I/O on the carrier board: 4x USB 3.2, 2x MIPI CSI camera connectors, Gigabit Ethernet, DisplayPort, 40-pin GPIO header and M.2 slots for NVMe SSD and wireless

Specifications

Selection facts.

AI Performance	Up to 67 TOPS (INT8) in Super mode
GPU	1024-core NVIDIA Ampere GPU with 32 Tensor Cores, up to 1020 MHz
CPU	6-core Arm Cortex-A78AE v8.2 64-bit, up to 1.7 GHz
Memory	8GB 128-bit LPDDR5, 102 GB/s
Storage	microSD slot plus M.2 Key M (NVMe SSD) slot on carrier board
Power	7W - 25W configurable (via included DC barrel jack or USB-C)
Camera	2x MIPI CSI-2 connectors (up to 4 cameras with virtual channels)
Networking / Expansion	Gigabit Ethernet, M.2 Key E for Wi-Fi/BT, 40-pin GPIO header
Video I/O	DisplayPort 1.2, 4x USB 3.2 Gen2 Type-A, USB-C
Software	NVIDIA JetPack SDK (Ubuntu-based Linux, CUDA, cuDNN, TensorRT, ROS 2 compatible)

Applications

Where it is used.

Running local generative AI and small language models (LLMs/SLMs) at the edge without cloud
Autonomous mobile robots and drones using ROS 2 for navigation and perception
Real-time computer vision: object detection (YOLO), segmentation and pose estimation
Smart cameras and video analytics for retail, traffic and security
Robotics and AI research, prototyping and university coursework
AI-powered inspection and quality control on production lines

Buying guide

How to choose.

Buy this kit if you have outgrown a Raspberry Pi for AI work and need real GPU-accelerated inference for vision, robotics or local LLMs. It suits robotics engineers, AI startups, final-year and research students, and anyone deploying neural networks where cloud latency or connectivity is a problem. If your project is only basic sensor reading or GPIO control, a microcontroller or Pi is more cost-effective.

The dev kit ships with the module already mounted on the reference carrier board. Plan to add a compatible NVMe M.2 SSD (strongly recommended over microSD for OS and model storage), an M.2 Wi-Fi/Bluetooth card if you need wireless, and a MIPI CSI camera (such as an IMX219 or IMX477 module) for vision projects. A quality 5V/DC supply meeting the 25W profile and active cooling (the kit includes a fan) are important for sustained Super-mode performance.

Pair it with JetPack, the NVIDIA Jetson community containers and ROS 2 for the fastest start. For higher performance or more memory later, the same carrier ecosystem scales up to the Orin NX modules.

FAQ

Common questions.

What is the difference between the 'Super' kit and the original Orin Nano dev kit?

The hardware is essentially the same Jetson Orin Nano 8GB module and carrier board. 'Super' refers to a free JetPack software update that adds a higher-clock power mode, raising AI performance from about 40 TOPS to up to 67 TOPS and increasing memory bandwidth. Existing original kits can be updated to Super performance.

Does it come with storage and an operating system?

The kit does not include a microSD card or SSD. You flash NVIDIA JetPack (an Ubuntu-based Linux with CUDA/TensorRT) onto a microSD card or, better, an NVMe M.2 SSD that you add to the M.2 slot. A 64GB or larger card/SSD is recommended.

Can it run local LLMs like Llama?

Yes. With 8GB LPDDR5 and Tensor Cores it can run quantized small language models and vision-language models locally. NVIDIA provides tutorials and containers (jetson-ai-lab) for running LLMs, SLMs and RAG on the device, though very large models will need quantization.

What camera and accessories do I need?

For vision, use a MIPI CSI-2 camera such as a Raspberry Pi IMX219 or high-res IMX477 module, or a USB webcam. For wireless, add an M.2 Key E Wi-Fi/Bluetooth card. An NVMe SSD, a compatible power supply and the included fan/heatsink complete a typical setup.

Is it available in India and is it easy to program?

Yes, it is stocked by Indian robotics and electronics distributors including BotBit. Development uses standard Python, C++, CUDA, PyTorch/TensorFlow and ROS 2, all supported by the JetPack SDK, so most existing NVIDIA and Jetson tutorials apply directly.

What is in the box?

The Jetson Orin Nano 8GB module pre-installed on the reference carrier board, with heatsink and fan. Power adapter, storage, camera and Wi-Fi module are typically purchased separately depending on the seller's bundle.