NVIDIA announces Jetson Orin Nano 2, a robotics computer aimed at entry-level edge AI

NVIDIA on August 25, 2026 announced the Jetson Orin Nano 2, a new robotics computer the company says brings frontier-class generative AI performance to entry-level edge devices. "NVIDIA today announced NVIDIA Jetson Orin Nano 2, a new robotics computer set to redefine entry-level edge AI — putting frontier-class generative AI performance in the hands of millions of developers," the company said in its announcement.

What's new

The Jetson Orin Nano 2 packs 78 trillion operations per second (TOPS) of AI compute, 8GB of memory, and an 8-core Arm CPU into the same compact form factor as its predecessor, the Jetson Orin Nano Super. NVIDIA says the new module delivers twice the inference performance of that predecessor while consuming 40% less power at the same performance level when run in its 15-watt mode.

NVIDIA is positioning the module for robots, delivery and inspection drones, and vision AI systems — the kind of cost- and power-constrained devices that sit below its higher-end Jetson Thor and Jetson AGX Orin lines. The module and developer kit are expected to be available in the first half of 2027; NVIDIA did not disclose pricing.

Deepu Talla, NVIDIA's VP of Robotics and Edge AI, framed the release around the shrinking gap between small and large models: "Today's small and medium frontier models have reached the accuracy of last year's largest frontier models, unlocking real-time intelligence for edge devices. The Jetson Orin Nano 2 computer puts that breakthrough within reach of millions of developers, delivering the performance and energy efficiency needed for real-time reasoning at the edge."

NVIDIA says more than 3 million developers have already built on its robotics stack, and named Cognex, Doosan Bobcat, and Matic among the first companies adopting and evaluating the new module.

Context

The announcement lands the same week NVIDIA pushed hardware news across several fronts: the Groq 3 LPX inference accelerator entered full production, the Vera Rubin NVL72 rack was pitched at 30x more agentic AI throughput per watt than the prior GB300 generation, and NVLink Fusion opened to custom XPUs for large AI factories. Jetson Orin Nano 2 targets the opposite end of that spectrum — small, low-power, on-device hardware rather than data-center-scale racks — but the underlying story is the same one NVIDIA has been telling all year: newer, smaller models are now capable enough to run useful generative AI locally, and NVIDIA wants dedicated silicon at every point on that spectrum.

Why it matters

Edge robotics has historically forced a tradeoff between running a capable model and fitting it into a device's power and cost budget. NVIDIA's claim that Jetson Orin Nano 2 delivers double the inference performance of its predecessor at a lower power draw, in an unchanged form factor, matters because it lets existing robot and drone designs upgrade their on-board intelligence without a hardware redesign. With a 2027 H1 availability window, the near-term impact is limited to companies planning next-generation edge products now, but the 3-million-developer robotics-stack figure NVIDIA cites suggests a large existing base of designs that could adopt the new module as a drop-in upgrade once it ships.