New wireless modules combine Bluetooth LE, multi-protocol connectivity, expanded memory, and local AI acceleration to simplify smarter, low-power embedded devices.
Ezurio has launched its BL54LM20 Series, a new family of wireless modules designed to add edge AI and multi-protocol connectivity to Bluetooth LE products. The series includes the BL54LM20A and BL54LM20B modules, along with a BL54LM20B USB-C Adapter for development and evaluation.
The BL54LM20B is the higher-performance option, integrating an Axon neural processing unit (NPU) for local machine-learning inference. According to Ezurio, the NPU can provide up to 15× faster AI inference than CPU-only processing, allowing connected devices to process sensor data locally rather than continuously sending it to the cloud.
Both modules are based on Nordic Semiconductor’s nRF54LM20A/nRF54LM20B SoCs and combine a 128MHz Arm Cortex-M33 application processor with a 128MHz RISC-V coprocessor. They provide 2MB non-volatile memory and 512KB RAM, giving developers additional headroom for wireless stacks, application firmware, sensor processing, and AI workloads.
The key features are:
- 2.4GHz multiprotocol wireless operation
- Integrated NFC interface
- USB connectivity for embedded designs
- Chip antenna and MHF4 RF options
- -40°C to +85°C operating temperature range
For connectivity, the modules support Bluetooth LE, 802.15.4-based Thread and Matter, and NFC. This makes them suitable for products that need to operate across multiple wireless ecosystems without requiring separate connectivity solutions.The modules measure 11×8×1.8mm and provide interfaces including USB, UART, SPI, I2S, PWM, ADC, and GPIO. Designers can select versions with an integrated chip antenna or an MHF4 connector, depending on their RF implementation requirements.
The BL54LM20B’s local AI capability can enable applications such as predictive maintenance, anomaly detection, voice or sound recognition, wearable devices, asset tracking, robotics, industrial sensing, and smart-building equipment. Processing data at the edge can also reduce latency and limit the amount of sensor information that needs to be transmitted.
Another advantage is the common form factor between the BL54LM20A and BL54LM20B. Developers can therefore design around a common hardware footprint and move toward AI-enabled functionality without completely redesigning the wireless section.The modules are also supported by Nordic’s development ecosystem and edge-AI tools, helping developers bring trained machine-learning models into resource-constrained Bluetooth products.
Click here for the original announcement.
