NVIDIA Jetson Orin Nano 2

Models & Pricing

Specifications

Overview

The NVIDIA Jetson Orin Nano 2 is the new entry level of the Jetson Orin family, announced in August 2026 as a substantial generational step rather than a simple speed bump. Where the original Orin Nano shipped as a heavily de-featured cut of the AGX Orin die, Orin Nano 2 is built on an entirely new chip — still Ampere architecture and still Cortex-A78, but with architecturally improved Tensor Core paths that NVIDIA says double the real-world work extractable from the hardware, even though peak theoretical throughput rises only about 16%.

Key deltas versus the Orin Nano 8GB / Orin Nano Super: AI performance rises from 67 TOPS to 78 TOPS (INT8 sparse), CUDA cores from 1,024 to 1,536, CPU cores from 6 to 8 Cortex-A78AE, and memory bandwidth from 102.4 GB/s to 120 GB/s thanks to LPDDR5X-7500 support. The power envelope widens to 15W–40W, matching Orin NX, which gives designers the option to trade efficiency against headroom.

Critically for integrators, Orin Nano 2 keeps the same compact module mechanical envelope and 260-pin connector as the current Orin Nano platform. That means existing carrier boards designed around Orin Nano can be re-used, and the whole Orin software stack carries over — NVIDIA explicitly positions it as a drop-in upgrade path. Silicon and developer kits are expected in the first half of 2027, with Orin production committed through 2032 for long-lifecycle industrial programmes.

For robotics and vision AI teams that hit the ceiling of the original Nano, Orin Nano 2 is the natural stepping stone without jumping to Orin NX pricing or the considerably larger AGX Orin / Thor footprint.

Target Applications

Entry-level robotics, autonomous mobile robots, delivery and inspection drones, real-time vision AI systems, multi-camera machine vision, generative and multimodal AI at the edge, smart retail analytics, industrial inspection, and physical AI development programmes.

Related Products