Artificial intelligence is increasingly moving beyond cloud data centers and into the physical world. Robots, drones, and intelligent vision systems need to process information locally, make decisions quickly, and operate reliably even when connectivity is limited.
Against this backdrop, NVIDIA has announced the Jetson Orin Nano 2, a new robotics computer designed to bring advanced edge AI capabilities to entry-level physical AI applications. NVIDIA announced the product on August 25, 2026.
What Is NVIDIA Jetson Orin Nano 2?
The NVIDIA Jetson Orin Nano 2 is a compact robotics computer designed for edge AI applications, including robots, delivery and inspection drones, and vision AI systems.
Instead of sending every piece of sensor or camera data to a remote cloud server, edge AI allows devices to process information closer to where it is generated. This can help reduce latency and enable machines to make decisions in real time.
NVIDIA says the Jetson Orin Nano 2 is designed to put frontier-class generative AI capabilities within reach of a much broader developer community. More than 3 million developers are already building on NVIDIA’s robotics stack, according to the company.
Key Specifications of Jetson Orin Nano 2
The new platform offers a substantial increase in performance while maintaining a compact form factor.
According to NVIDIA, Jetson Orin Nano 2 features:
- 78 trillion operations per second (TOPS) of AI compute
- 8GB of memory
- 8-core Arm CPU
- Improved Tensor Cores
- Higher memory bandwidth
- Support for generative AI and vision-language models
- NVIDIA’s open software ecosystem
NVIDIA says the Jetson Orin Nano 2 delivers 2x the inference performance of the Jetson Orin Nano Super while retaining the same compact form factor. At a 15-watt operating mode, it can deliver the same performance as its predecessor while consuming 40% less power, according to NVIDIA.
Why Edge AI Matters
Traditional cloud-based AI can be extremely powerful, but applications that operate in the physical world often require immediate decision-making.
Consider an autonomous drone approaching an obstacle or a robot navigating a busy warehouse. Sending sensor information to the cloud and waiting for a response can introduce latency and create a dependency on network connectivity.
Edge AI addresses this by moving computation closer to the device.
Lower Latency
Local processing can allow robots and drones to respond more quickly to sensor information, which is particularly important for navigation, obstacle avoidance and real-time control.
Greater Reliability
Edge devices can continue processing AI workloads even when network connectivity is poor or unavailable. This is useful for applications such as remote infrastructure inspection, agriculture and disaster response.
Improved Data Privacy
Processing sensitive information locally can reduce the need to transmit certain data to remote servers. This can be particularly valuable for applications involving cameras, industrial environments and other sensitive information.
Reduced Cloud Dependency
Organizations can potentially reduce their reliance on cloud processing and data transfers by performing more AI inference directly on edge devices.
Generative AI Comes to Physical AI
One of the most important aspects of the Jetson Orin Nano 2 is its focus on running modern AI models directly on physical machines.
NVIDIA says the platform can run large language models and vision-language models optimized for memory-efficient edge inference. Supported open models highlighted by NVIDIA include NVIDIA Cosmos, NVIDIA Nemotron, Gemma 4 and Qwen 3.
This points toward a future where robots are not limited to traditional programmed responses but can use increasingly capable AI models to interpret their surroundings, understand instructions and respond to changing situations.
Applications in Robotics
The Jetson Orin Nano 2 could support a wide range of robotics applications.
Autonomous Mobile Robots
Robots operating in warehouses, factories and other environments can use AI for perception, mapping, navigation and decision-making.
Industrial Robotics
Edge AI can enable robots to analyze camera feeds in real time for quality inspection, object detection, assembly and other manufacturing tasks.
Home Robots
Compact computing hardware can help consumer robots process information locally for perception, navigation and interaction.
NVIDIA cited Matic as one of the early companies exploring the Jetson Orin Nano 2 for a compact home robotics platform focused on real-time perception, interaction and navigation.
Human-Robot Collaboration
More capable edge AI could help robots better understand their surroundings and operate alongside humans in shared environments.
Applications in Drones
Drones are another major target for the new platform.
Autonomous Inspection
Inspection drones can process visual information locally to identify potential problems in infrastructure, industrial facilities and other assets.
Delivery Drones
Edge computing can support navigation, obstacle detection and autonomous decision-making for delivery systems.
Agriculture
AI-powered drones can analyze crops and identify visual indicators of plant stress, pests or other agricultural conditions.
Search and Rescue
Local AI processing could help drones analyze large amounts of visual information while operating in areas where communications infrastructure may be limited.
NVIDIA’s Robotics Software Ecosystem
Hardware is only one part of an edge AI platform. NVIDIA’s software ecosystem is a major component of the Jetson platform.
The Jetson ecosystem includes technologies for accelerated AI, robotics and computer vision development.
NVIDIA’s robotics stack includes software such as JetPack, Isaac technologies and other development tools designed to help developers build and deploy AI applications.
The company also says Jetson Orin Nano 2 is powered by its open software stack and Jetson agent skills, providing developers with tools for building advanced AI applications at the edge.
Growing Partner Ecosystem
NVIDIA is also building a broader hardware and software ecosystem around Jetson Orin Nano 2.
The company identified Cognex, Doosan Bobcat, and Matic among the first companies adopting or exploring the platform.
Other ecosystem partners are developing carrier boards, customized hardware, AI software and reference solutions based on Jetson technology.
This ecosystem could help developers move from prototypes to production systems more quickly.
Jetson Orin Nano 2 Availability
The Jetson Orin Nano 2 module and developer kit are expected to become available in the first half of 2027, according to NVIDIA.
Therefore, although NVIDIA has announced the platform, developers should distinguish between the announcement date and actual commercial availability.
What Makes Jetson Orin Nano 2 Important?
The biggest significance of the Jetson Orin Nano 2 is not simply its computing performance. It is the broader shift toward physical AI.
As AI models become more efficient, powerful models can increasingly run on smaller devices. This creates opportunities to bring sophisticated AI capabilities into machines that previously relied on simpler algorithms or remote cloud services.
The combination of 78 TOPS of AI compute, 8GB memory, an 8-core Arm CPU, and improved inference performance gives developers a platform aimed at bringing these capabilities into relatively compact physical systems.
Jetson Orin Nano 2 vs. Traditional Cloud AI
The two approaches are not necessarily competitors. In many real-world systems, edge and cloud AI can work together.
A robot could process time-sensitive sensor information locally while sending selected information to the cloud for deeper analysis, training or centralized management.
This hybrid approach could provide a balance between real-time response, computing efficiency and access to larger cloud-based AI systems.
The Future of Physical AI
The development of platforms such as Jetson Orin Nano 2 highlights a broader trend in artificial intelligence: AI is moving from software applications on computers and smartphones into machines that interact directly with the physical world.
Robots need to see and understand their surroundings. Drones need to navigate independently. Industrial machines need to detect defects. Autonomous systems need to make decisions without waiting for instructions from a remote server.
Edge AI hardware provides the computational foundation for these capabilities.
Summary to consider
NVIDIA’s Jetson Orin Nano 2 represents a significant step in bringing advanced AI inference to compact robotics and edge devices. With 78 TOPS of AI compute, 8GB memory, an 8-core Arm CPU and claimed 2x inference performance over Jetson Orin Nano Super, the platform is designed to make sophisticated edge AI more accessible to developers.
Its focus on robots, drones and vision AI systems also reflects NVIDIA’s broader push into physical AI, where intelligent machines can perceive, reason and act in the real world.
With commercial availability expected in the first half of 2027, the Jetson Orin Nano 2 could become an important development platform for the next generation of autonomous robots, drones and intelligent edge systems.
Disclaimer: Specifications, performance comparisons and availability mentioned in this article are based on NVIDIA’s August 25, 2026 announcement and may change before commercial release.