NVIDIA

Robotics · Edge AI · Semiconductors

Launch

NVIDIA launches Jetson Orin Nano 2, doubling edge AI performance

Shrinking frontier models now run on hardware cheap enough to put real-time reasoning inside drones and delivery robots.

• NVIDIA announced the Jetson Orin Nano 2 robotics computer on August 25, 2026, delivering 78 trillion operations per second of AI compute with 2x the inference performance of its predecessor.
• Jetson is NVIDIA's line of compact edge-computing modules for robots, drones and vision systems, with more than 3 million developers already building on its robotics software stack.
• The module keeps its predecessor's 8GB of memory and 8-core Arm CPU but adds improved Tensor Cores and higher memory bandwidth, cutting power draw by 40% at matched performance in 15-watt mode.
• The module and developer kit are expected to ship in the first half of 2027, with pricing not yet disclosed.
• Early adopters include Cognex, Doosan Bobcat and Matic Robots, while Alphabet's drone delivery unit Wing plans to evaluate the chip for real-time perception in its delivery fleet.
• The current Jetson Orin Nano Super tops out at 67 TOPS starting at $199, positioning Nano 2 as an upgrade to NVIDIA's entry tier below its pricier Orin NX, AGX Orin and Jetson Thor modules.
• As small and medium AI models now match the accuracy of last year's largest models, NVIDIA is betting that entry-level edge hardware can replace cloud-dependent processing for real-time robotics and drones.