Edge AI - Nordic Semiconductor
Edge AI helps IoT companies and device manufacturers implement on-device machine learning. It is designed for teams looking to reduce radio transmission frequency to support longer battery life.
At a glance
- Category
- Software Development
- Best for
- IoT Companies, Device Manufacturers, Software Companies building for embedded hardware
- Pricing
- Pricing was not clearly available from the provided evidence. Buyers should confirm current pricing on the vendor website.
- Key use cases
- Time-Series Sensor Analysis, Low-Power Wireless Communication, Audio and Image Classification, Anomaly Detection
- Official website
- nordicsemi.com
Edge AI is a development ecosystem from Nordic Semiconductor for running neural networks on wireless System-on-Chips (SoCs) and System-in-Packages (SiPs). The technology is designed to move data processing from the cloud to the device, which may help reduce latency and network traffic.
The ecosystem is for software companies and device manufacturers building ultra-low-power IoT hardware. It provides two paths for implementation: ultra-tiny Neuton models for CPU execution and the Axon NPU for hardware-accelerated TensorFlow Lite models.
By processing sensor data locally, the technology supports devices that need to operate on small batteries. It is specifically for those working within the Nordic hardware ecosystem, including the nRF series.
Buyers should confirm which specific Nordic SoC they are using, as the Axon NPU acceleration is only available on select hardware such as the nRF54LM20B.
Key Features
Tiny AI models with a memory footprint under 5KB designed to run on Nordic SoC CPU cores.
A dedicated hardware accelerator for TensorFlow Lite models designed for faster inference than a standard CPU.
A software package for the nRF Connect SDK that includes drivers, an Axon NPU compiler, and the nRF Edge AI Library.
A tool for creating embedded ML models based on the Edge Impulse platform.
Supports development for nRF52, nRF53, nRF54L, nRF70, and nRF91 Series devices.
Use Cases
On-device inference for data from accelerometers, IMUs, and temperature sensors.
Implementing AI within Bluetooth LE, Zigbee, Thread, and Matter protocols to help reduce radio transmission.
Using the Axon NPU for workloads such as keyword recognition or image classification tasks.
Processing data locally to detect patterns in device behavior.
Best For
Pricing
Pricing was not clearly available from the provided evidence. Buyers should confirm current pricing on the vendor website.
FAQ
Neuton models are ultra-tiny models designed to run on a standard CPU core, while the Axon NPU is a dedicated hardware accelerator for TensorFlow Lite models.
These tools are designed for Nordic Semiconductor hardware, including various nRF Series SoCs and SiPs.
They support time-series sensor data from accelerometers, IMUs, and temperature sensors, as well as audio and image classification.
Source category: Software Development
Source subcategory: AI Development Platform
