u-blox introduces ALMA-B2 modules with low-latency Edge ML capabilities
u-blox has announced the expansion of its short-range radio portfolio with new low-latency Edge Machine Learning (Edge ML) capabilities. The launch strengthens the collaboration between u-blox and Nordic Semiconductor with the introduction of the ALMA-B2 Bluetooth LE module series. Based on Nordic’s next-generation nRF54LM20 platform, ALMA-B2 combines next-generation Bluetooth LE connectivity with a powerful wireless MCU, delivering Channel Sounding, global certification, and flexible antenna options in a compact form factor, enabling a new class of intelligent, low-power IoT applications.
As IoT devices evolve to require faster response times, lower power consumption, and increased local intelligence, developers are shifting toward processing data at the edge rather than relying on cloud-based systems. The ALMA-B2, u-blox’s first module with a dedicated hardware Neural Processing Unit (NPU), addresses this demand by integrating Bluetooth LE 6.0 connectivity with a powerful wireless MCU architecture from Nordic Semiconductor, optimised for low-latency operation and advanced edge processing, making it up to 15x faster, with superior energy efficiency. This makes it ideal for applications such as healthcare, industrial automation and robotics, asset tracking, and smart home devices. Support for Bluetooth Channel Sounding enables highly accurate distance measurement, while professional-grade module options, global certification, and flexible antenna designs help manufacturers accelerate development cycles and streamline global product launches.
u-blox’s ALMA-B2 integrates a 128 MHz Arm Cortex-M33 processor, 2 MB of non-volatile memory and 512 kB of RAM, and a RISC-V coprocessor. ALMA-B21 variants come with a built-in Axon NPU enabling hardware-based machine learning capabilities. All module variants support Bluetooth LE 6.0 alongside 802.15.4 technologies such as Thread and Zigbee, and include a rich set of interfaces including; HS-USB for demanding embedded designs. Advanced security features such as TrustZone, cryptographic accelerators, tamper detection, debug port protection, and an immutable boot partition support, help developers create and build secure and robust connected products. With compact dimensions, an operating range of -40 °C to 85 °C, and both embedded and external antenna options, the ALMA-B2 provides OEMs with a flexible platform for next-generation IoT solutions. Built for lower power consumption and ultra-low latency, delivering up to 15x improvement compared to running algorithms solely on the MCU, the module leverages its NPU architecture to support complex applications efficiently.
