Synaptics and Google unveil open-source edge AI microcontrollers

Synaptics and Google have introduced a new series of low-power edge AI microcontrollers, including the Astra SR200 and SL2610/SL2600 series, designed for battery-powered devices such as wearables, smart glasses, and industrial sensors.

The hardware integrates Google’s open-source Coral NPU with Synaptics’ Torq AI engine, Arm Cortex-M CPU cores, and multimedia processing capabilities. This combination allows for simultaneous multimodal processing of voice, vision, and motion. A key technical differentiator is the integration of high-bandwidth DRAM that scales to hundreds of megabytes, allowing these microcontrollers to handle workloads that previously required Linux-based application processors.

The platform utilizes an open-source software stack based on MLIR/IREE compilers to reduce fragmentation in IoT edge computing. Performance metrics indicate that average on-device inference draws less than 6 milliwatts, with the Coral NPU reportedly delivering approximately 40% lower energy use compared to existing edge AI processors.