SemiconductorsBRIEF

NVIDIA Launches Jetson Orin Nano 2 Robotics Computer, Targets Entry-Level Edge AI

N
NathanTechnology Editor · Technical Lead
Published · Updated
NVIDIA announced the Jetson Orin Nano 2 robotics computer on August 25, 2026, packing 78 TOPS of AI compute, 8GB of memory, and an 8-core Arm CPU while doubling inference performance and cutting power draw 40% versus its predecessor. Modules and developer kits ship in the first half of 2027, targeting a robotics developer base NVIDIA says already tops 3 million.

When Is NVIDIA Launching the Jetson Orin Nano 2 Robotics Computer?

NVIDIA announced the Jetson Orin Nano 2 robotics computer on August 25, 2026, positioning it to redefine entry-level edge AICITE:E1. NVIDIA says the device is meant to give millions of developers access to frontier-level generative AI performanceCITE:E1. The launch materials describe the computer as letting developers build next-generation robots, delivery and inspection drones, and vision AI systemsCITE:E9. NVIDIA Jetson Orin Nano 2 modules and developer kits are expected to ship in the first half of 2027CITE:E8.

What Hardware Specs Define This Entry-Level Jetson?

The Jetson Orin Nano 2 delivers 78 TOPS of AI compute, 8GB of memory, and an 8-core Arm CPU in the same compact form factor as its predecessorCITE:E5CITE:E11. The module adds improved Tensor Cores and higher memory bandwidth over the prior generationCITE:E11. NVIDIA says the design balances cost-effectiveness and energy efficiency while substantially raising AI and video-processing performanceCITE:E5.

How Much Faster and More Efficient Is Inference on the New Chip?

NVIDIA says the Jetson Orin Nano 2 doubles inference performance versus the previous generation while cutting power consumption by 40% at equivalent performanceCITE:E2. The 2x inference gain is measured specifically against the Jetson Orin Nano Super, and NVIDIA attributes it to the improved Tensor Cores and higher memory bandwidthCITE:E11.

MetricValueEvidence
AI compute78 TOPSCITE:E5
Memory8GBCITE:E5
CPU8-core ArmCITE:E5
Inference performance vs. Jetson Orin Nano Super2xCITE:E2
Power draw at equal performance-40%CITE:E2
Robotics developer baseOver 3,000,000CITE:E3
Announcement dateAugust 25, 2026CITE:E1
Module/dev kit availabilityFirst half of 2027CITE:E8

How Large Is the Developer Ecosystem Building on NVIDIA's Robotics Stack?

NVIDIA states more than 3 million developers are already building on its robotics technology stackCITE:E3CITE:E12. Cognex, Doosan Bobcat, and Matic are named as early adopters exploring the Jetson Orin Nano 2CITE:E3. NVIDIA also names Wing alongside Cognex, Doosan Bobcat, and Matic as companies bringing edge AI applications — home robots, vision AI systems, delivery and inspection drones, carrier boards, hardware systems, and reference solutions — to compact devicesCITE:E12.

Which Companies Are Already Applying Jetson Orin Nano 2 to Real Products?

Alphabet's drone-delivery unit Wing and home-robot maker Matic both point to concrete uses for the new chipCITE:E6CITE:E7. Wing currently runs the Jetson Orin Nano Super and NVIDIA's software stack across its delivery-drone fleet, and plans to evaluate the Jetson Orin Nano 2 to further improve real-time AI perception and inference for deliveries ranging from local stores to home yardsCITE:E6. Matic co-founder Navneet Dalal says the chip lets Matic run state-of-the-art AI models at the edge on a compact home-robot platform built for real-time perception, interaction, and navigationCITE:E7.

What Does NVIDIA Say About the Outlook for Edge AI and Robotics?

NVIDIA robotics and edge AI vice president Deepu Talla says mid-size and small frontier models now match the accuracy of last year's largest frontier modelsCITE:E4CITE:E10. Talla says this opens new possibilities for real-time intelligence on edge devices, and that the Jetson Orin Nano 2 brings that breakthrough to millions of developers with the performance and energy efficiency needed for real-time inference at the edgeCITE:E4CITE:E10.

Taken together, the spec sheet — 78 TOPS, 8GB, 8-core Arm, a 2x inference gain, and a 40% power cut at equal performance in an unchanged form factor — is paired with a developer base NVIDIA puts above 3 million and named early adopters spanning drone delivery (Wing), home robotics (Matic), and industrial vision and equipment (Cognex, Doosan Bobcat). The module and developer kit do not ship until the first half of 2027, so the adopters cited so far are describing evaluation and fleet use of the platform rather than shipped end products.

📊 Evidence

FAQ

When Is NVIDIA Launching the Jetson Orin Nano 2 Robotics Computer?

NVIDIA announced the Jetson Orin Nano 2 robotics computer on August 25, 2026, positioning it to redefine entry-level edge AICITE:E1.

What Hardware Specs Define This Entry-Level Jetson?

The Jetson Orin Nano 2 delivers 78 TOPS of AI compute, 8GB of memory, and an 8-core Arm CPU in the same compact form factor as its predecessorCITE:E5CITE:E11.

How Much Faster and More Efficient Is Inference on the New Chip?

NVIDIA says the Jetson Orin Nano 2 doubles inference performance versus the previous generation while cutting power consumption by 40% at equivalent performance…

How Large Is the Developer Ecosystem Building on NVIDIA's Robotics Stack?

NVIDIA states more than 3 million developers are already building on its robotics technology stackCITE:E3CITE:E12.

📎 Sources

  1. blogs.nvidia.com.tw
  2. cna.com.tw

Related data

Author's TakeNathan

The headline pairing here isn't raw performance but performance per watt: a 2x inference gain combined with a 40% power cut at equal performance, delivered in the same form factor as the prior Jetson Orin Nano. That combination matters more for entry-level edge AI than peak TOPS, because compact platforms like Matic's home robot and Wing's delivery drones are bounded by battery and thermal budgets, not by available compute headroom. With modules and dev kits not shipping until the first half of 2027, the metric worth tracking isn't NVIDIA's 3-million-developer stack figure — it's whether the named early adopters, Wing, Matic, Cognex, and Doosan Bobcat, convert their current evaluation and fleet use of Jetson Orin Nano 2 (and its predecessor, Jetson Orin Nano Super) into shipped, fielded products.

N
NathanTechnology Editor · Technical Lead

Related

BRIEF

SpaceX and NVIDIA Plan to Launch Orbital AI Data Centers Starting Next Year, Musk Says

SpaceX plans to launch its first NVIDIA-chip-powered AI satellites, Starmind AI1, in the fourth quarter of 2027, reaching large-scale deployment by 2028 — at least a year ahead of the original schedule. The company has also filed with the FCC for a network of up to 1 million satellites, while Taiwanese suppliers Unitech and Sesoda report rising satellite-related revenue and order shares tied to the buildout.

林紀旭 James Lin ·
BRIEF

HanTest's January–July 2026 Revenue Already Tops Full-Year 2025 Ahead of TPEx Listing

HanTest (漢測) reported NT$2.668 billion in cumulative revenue for January–July 2026, up 128.18% year-on-year and already exceeding its entire 2025 revenue of NT$2.425 billion, as the company prepares for a late-September 2026 Taipei Exchange listing built on expansion into memory testing, advanced packaging, and silicon photonics.

EffectStory 編輯部 ·