Nordic Semiconductor has introduced its nRF71 Series of wireless systems-on-chip (SoCs), combining Wi-Fi, Bluetooth Low Energy, Thread connectivity, and hardware-accelerated edge artificial intelligence on a single platform. Designed for battery-powered connected devices, the new processors target applications including smart home equipment, wearables, security systems, and industrial Internet of Things (IoT) sensors. The launch addresses a growing challenge in edge AI development: delivering local machine learning capabilities without significantly increasing power consumption. Nordic says the nRF71 Series delivers up to 20 times lower standby power consumption than its previous-generation nRF70 Wi-Fi products, potentially expanding the range of devices capable of running AI inference while operating on batteries.
3. Article
Nordic Semiconductor Combines Wi-Fi and Edge AI in nRF71
Nordic Semiconductor is expanding its wireless computing portfolio with the nRF71 Series, a new family of ultra-low-power SoCs designed to bring connectivity and artificial intelligence processing closer to the devices generating data.
The platform combines Wi-Fi, Bluetooth Low Energy (LE), Thread, and edge AI capabilities within a single chip. Rather than relying on separate wireless modules and external AI processors, developers can use the integrated architecture to build connected devices that process information locally while maintaining multiple wireless connections.
Power consumption is central to the announcement. Nordic claims the nRF71 Series achieves up to 20 times lower standby power usage than its current-generation nRF70 Series Wi-Fi products. The comparison concerns standby power rather than overall device energy consumption, which will also depend on workloads, connectivity patterns, and hardware configuration.
The development reflects an important direction in edge computing: moving more intelligence onto compact, battery-powered devices instead of transmitting every data point to cloud-based AI services.
Integrated NPU Accelerates Machine Learning Inference
One of the platform’s major features is Nordic’s Axon Neural Processing Unit (NPU), which provides dedicated hardware acceleration for machine learning workloads.
According to the company, the NPU can execute machine learning models between 10 and 17 times faster than the 256 MHz Arm Cortex-M33 processor operating alone.
This hardware acceleration is intended to support applications such as keyword spotting, anomaly detection, and image classification. These workloads are relevant to devices that need to recognize patterns, respond to events, or make decisions without continuously communicating with a remote server.
The nRF71 Series also supports smaller machine learning models developed using Nordic’s Neuton software. These models can run directly on the processor without requiring the Axon NPU, giving developers another option when balancing inference performance, memory requirements, and power consumption.
Nordic is combining this processing architecture with nRF Cloud for over-the-air model updates. This allows developers to update deployed machine learning models as application requirements evolve, rather than requiring physical device replacement.
The combination of wireless connectivity, application processing, and AI acceleration on one chip could reduce component requirements and simplify hardware design. However, practical performance will depend on model complexity, available memory, and the specific inference workload.
1. SEO Headline
Nordic Unveils nRF71 Chips to Bring AI to Wi-Fi IoT
TechEdgeAI | AI Chips & Infrastructure | Edge AI | Semiconductor Technology
2. Opening Paragraph
Nordic Semiconductor has introduced its nRF71 Series of ultra-low-power wireless systems-on-chip (SoCs), combining Wi-Fi, Bluetooth Low Energy (LE), Thread connectivity, and edge artificial intelligence (AI) capabilities for battery-powered connected devices. The new chips target smart home products, wearables, security systems, and industrial Internet of Things (IoT) applications that need local AI processing without sacrificing battery life. Nordic claims the platform delivers up to 20 times lower standby power consumption than its previous-generation nRF70 Wi-Fi products. By integrating wireless connectivity and machine learning acceleration into a compact computing platform, the company is targeting a longstanding limitation in edge AI: running intelligent applications on devices that cannot depend on frequent charging or continuous access to cloud computing.
Nordic Semiconductor Targets Battery-Powered Edge AI With nRF71
Nordic Semiconductor is expanding its low-power wireless computing portfolio with the nRF71 Series, a new family of SoCs designed to bring AI inference and multiple wireless technologies into battery-powered devices.
The announcement addresses an important engineering challenge for edge AI developers. While machine learning models are increasingly capable of processing information locally, adding AI functionality to small connected devices can increase processor workloads, memory requirements, and power consumption. Wi-Fi introduces another constraint because maintaining wireless connectivity can consume a substantial portion of a device’s energy budget.
Nordic says the nRF71 Series delivers up to 20 times lower standby power consumption than its current-generation nRF70 Series Wi-Fi products. The comparison is specifically about standby power, not total operating energy, which will vary with network activity, processing workloads, and device configuration.
The platform combines Wi-Fi, Bluetooth LE, Thread, and application processing, with dedicated hardware acceleration for machine learning. Its objective is to reduce the number of separate components required to build intelligent connected products.
Axon NPU Brings Machine Learning Acceleration On-Device
A central component of the new platform is Nordic’s Axon Neural Processing Unit (NPU), designed to accelerate machine learning inference directly on the chip.
According to Nordic, the NPU executes machine learning models between 10 and 17 times faster than the 256 MHz Arm Cortex-M33 processor operating alone.
This acceleration targets practical edge AI workloads, including keyword spotting, anomaly detection, and image classification. These applications allow devices to recognize voice commands, identify unusual operating patterns, or classify visual information without sending every input to a remote AI service.
The architecture could be particularly useful in industrial sensors and security equipment, where local processing can reduce response times and dependence on network availability.
Nordic is also supporting smaller machine learning models through its Neuton software. Developers can train models using their own data and deploy ultra-small models directly on the processor, without necessarily requiring dedicated NPU acceleration.
The company is extending this development environment through nRF Cloud, which supports over-the-air model updates. This gives manufacturers a way to update deployed AI models as device requirements and operating conditions change.
The combination of an integrated NPU, embedded processing, and cloud-supported model management positions the nRF71 Series as more than a conventional wireless connectivity chip. It is designed to provide a complete foundation for low-power intelligent IoT applications.
Wi-Fi, Bluetooth and Thread Operate Simultaneously
Another important feature is the platform’s wireless architecture.
Nordic has incorporated two independent radio systems with separate radio-frequency and processing blocks. One handles Wi-Fi, while the other uses the 2.4 GHz multiprotocol radio technology from its latest nRF54 Series.
Both radios can receive simultaneously. This allows a connected device to maintain its Wi-Fi connection to cloud services while listening for Bluetooth LE or Thread communications without requiring one protocol to wait for the other.
That capability is relevant to Matter-enabled smart home products, including smart locks, security panels, hubs, bridges, and Thread border routers. Industrial sensors and wearables are also among the intended applications.
The design could simplify products that previously required multiple wireless components to maintain connections across different protocols.
nRF7120 and nRF7120E Introduce Wi-Fi 6 and Wi-Fi 6E
The first products in the series are the nRF7120 and nRF7120E, with different wireless capabilities.
- nRF7120: Supports dual-band Wi-Fi 6.
- nRF7120E: Adds tri-band Wi-Fi 6E, including the 6 GHz band, alongside hardware-accelerated edge AI capabilities.
Both chips include 4 MB of on-chip non-volatile memory and 1 MB of user-accessible RAM.
The memory configuration is important for embedded AI development because inference models, application code, and wireless communication functions must operate within relatively constrained hardware resources.
Nordic is building on its existing Bluetooth LE development ecosystem, including software, tools, and wireless technologies. This could lower the development barrier for manufacturers already using the company’s nRF platforms who want to introduce Wi-Fi and local AI functionality.
The products are currently sampling with selected customers, rather than being generally available for volume production. Nordic has not disclosed commercial pricing or a broader production availability schedule.
Edge AI Adoption Expands Beyond Smartphones and PCs
The launch reflects a broader transition in AI computing. Instead of relying exclusively on centralized cloud platforms or powerful local computers, developers are increasingly embedding machine learning capabilities into sensors, appliances, industrial equipment, and other connected devices.
According to Omdia’s October 2026 research, 62% of surveyed organizations have adopted edge computing for IoT deployments, while 42% identify edge processing as their leading technology investment priority. The research surveyed 570 IoT decision-makers across 10 countries. Omdia
Gartner’s 2026 Hype Cycle for Edge Computing also identifies AI as a major accelerator of edge computing innovation, particularly for applications interacting with the physical world. Gartner
This market includes established semiconductor suppliers such as Qualcomm, NXP Semiconductors, and STMicroelectronics, which are developing processing platforms for connected and intelligent devices. Nordic’s differentiation is its emphasis on combining low-power wireless connectivity with embedded AI capabilities.
The competitive opportunity is not limited to raw AI processing speed. Battery life, connectivity reliability, development tools, memory capacity, and long-term software support are equally important when manufacturers select silicon for products expected to operate for years.
What the nRF71 Series Means for IoT Developers
For manufacturers, combining Wi-Fi, Bluetooth LE, Thread, and AI processing could simplify hardware design and reduce component requirements. A smart lock, industrial sensor, or wearable device could potentially perform local inference while maintaining multiple wireless connections through one integrated platform.
There are also potential advantages in data handling. Local inference can reduce the need to transmit raw sensor information continuously, potentially lowering network traffic and cloud processing requirements. It can also help devices respond when connectivity is intermittent.
However, actual battery-life improvements will depend on how frequently devices communicate, how complex their AI models are, and how often they perform inference. The 20-times standby-power claim should not be interpreted as a universal improvement in overall battery life.
Nordic’s next challenge will be demonstrating how the platform performs in commercial products, particularly when developers combine continuous wireless connectivity with real-world machine learning workloads.
The nRF71 Series signals a growing convergence between semiconductor design, wireless infrastructure, and edge AI. As intelligent functionality moves into smaller devices, energy-efficient processing may become just as important as model performance in determining where AI can be deployed.
Market Landscape
The edge AI semiconductor market is evolving as manufacturers look to integrate machine learning into connected devices without relying on continuous cloud connectivity.
Nordic’s announcement reflects five important industry developments:
- Low-power AI chips: Semiconductor companies are developing dedicated NPUs to accelerate inference while controlling energy consumption.
- Wireless convergence: Wi-Fi, Bluetooth LE, and Thread integration is simplifying connected-device architecture.
- On-device intelligence: Local inference is expanding across smart homes, industrial automation, security, and wearable technology.
- AI-enabled IoT infrastructure: Edge computing is moving from experimental deployments toward wider commercial implementation.
- Model lifecycle management: Over-the-air AI model updates are becoming important for maintaining intelligent devices after deployment.
Nordic competes in a broader semiconductor ecosystem that includes Qualcomm, NXP Semiconductors, STMicroelectronics, and other suppliers of embedded processing and wireless connectivity technologies.
Its differentiation is the combination of ultra-low-power wireless connectivity, embedded machine learning acceleration, and developer software within one platform.
Top Insights
- Nordic’s nRF71 Series combines Wi-Fi, Bluetooth LE, Thread, and edge AI capabilities to simplify battery-powered intelligent IoT device development.
- The integrated Axon NPU accelerates machine learning inference between 10 and 17 times compared with the 256 MHz Cortex-M33 processor alone.
- Nordic claims up to 20 times lower standby power consumption compared with its previous-generation nRF70 Wi-Fi products.
- The nRF7120 supports dual-band Wi-Fi 6, while the nRF7120E adds tri-band Wi-Fi 6E and hardware-accelerated edge AI.
- The platform targets smart home, industrial IoT, wearables, and security applications requiring local intelligence and energy-efficient wireless connectivity.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
