Updated
Updated · NVIDIA Blog · Aug 25
NVIDIA Unveils 78-TOPS Jetson Orin Nano 2, Doubling Edge AI Performance
Updated
Updated · NVIDIA Blog · Aug 25

NVIDIA Unveils 78-TOPS Jetson Orin Nano 2, Doubling Edge AI Performance

3 articles · Updated · NVIDIA Blog · Aug 25

Summary

  • Jetson Orin Nano 2 packs 78 trillion operations per second, 8GB of memory and an 8-core Arm CPU for entry-level robots, drones and vision AI systems, with module and developer kit availability slated for first-half 2027.
  • NVIDIA said the new computer delivers 2x the inference performance of Jetson Orin Nano Super in the same form factor and uses 40% less power in 15-watt mode for equivalent performance.
  • More than 3 million developers already build on NVIDIA’s robotics stack, and the system is designed to run memory-efficient LLMs and vision-language models including Cosmos, Nemotron, Gemma 4 and Qwen 3 at the edge.
  • Cognex, Doosan Bobcat and Matic are among early adopters, while Wing plans to evaluate the platform for faster, safer drone delivery and partners are building carrier boards, hardware systems and reference designs to speed deployment.

Insights

Jetson Orin Nano 2 promises 2x inference and lower power, but will software, firmware, and deployment complexity decide its real-world success?
Why is NVIDIA betting that smaller, lower-power edge computers—not bigger cloud systems—will define the next wave of physical AI?