NVIDIA has announced the launch of the Jetson Orin Nano 2, a compact edge robotics computer engineered to bring physical AI and real-time generative intelligence to autonomous machines, delivery drones, and advanced vision systems. The board is designed as an entry-level hardware solution for developers looking to execute sophisticated AI workloads directly on edge devices rather than relying on cloud data centres.
According to NVIDIA, the launch aligns with rapid advancements in small and medium-scale AI models, which have evolved to deliver accuracy levels previously restricted to massive data-centre-scale frontier architectures. This efficiency allows compact, low-power systems to handle complex vision and language processing on the fly.
Compute Capabilities and Energy Efficiency
The Jetson Orin Nano 2 integrates an eight-core Arm CPU, 8GB of memory, and delivers up to 78 trillion operations per second (TOPS) of AI compute performance. While retaining the compact form factor of its predecessor, the board achieves double the inference performance of the Jetson Orin Nano Super, driven by upgraded Tensor Cores and expanded memory bandwidth.
Energy efficiency is a key highlight of the new architecture: when configured in a 15-watt operating mode, the board cuts power consumption by 40 percent compared to the Orin Nano Super while maintaining an equivalent performance benchmark.
The module is fully integrated into NVIDIA’s software ecosystem, supporting large language models and vision language models that are specifically optimised for memory-constrained edge deployment. Supported architectures include NVIDIA’s own Nemotron and Cosmos models, alongside open-weight options such as Gemma 4 and Qwen 3.
Adoption in Delivery Drones and Consumer Robotics
Multiple industry partners have begun integrating or evaluating the Jetson Orin Nano 2 across various robotics and automation sectors:
Wing: The Alphabet-owned autonomous drone delivery service, which currently uses the Jetson Orin Nano Super, is evaluating the new board to enhance onboard spatial perception and real-time decision-making for commercial yard deliveries.
Matic Robots: The domestic robotics manufacturer is implementing the processor in its autonomous home cleaning machines to drive semantic environment mapping, gesture recognition, and natural language communication.
Industrial Automation: Companies such as Cognex and Doosan Bobcat are exploring the platform for edge-based vision and machinery control applications.
Hardware Ecosystem and Carrier Solutions
To support commercial deployment across various industries, NVIDIA is collaborating with a wide network of hardware manufacturers and system integrators. Partners including AAEON, ADLINK, Advantech, Aetina, Connect Tech, and Seeed Studio are developing custom carrier boards, reference designs, and ready-to-deploy hardware solutions around the Jetson Orin Nano 2 platform.
Comments
to leave a comment.
No comments yet.