Ceva, Inc. NPU family debuts in Mobilint's REGULUS SoC

Ceva, Inc. reported that Mobilint has placed components from Ceva's NeuPro NPU family into production for the first time. The development matters because AI workloads are moving onto power-constrained edge devices rather than running only in the cloud.

Workloads are being assigned to whichever location runs them most effectively, from the cloud to the device itself. For Physical AI systems that must sense, interpret and respond to their surroundings in real time, that shift pushes greater intelligence directly into the device.

The REGULUS edge AI system-on-chip is designed as a standalone processor for on-device AI. It integrates Ceva's SensPro AI DSP and NeuPro-Nano NPU within Mobilint's processing architecture, delivers 10 TOPS of AI performance at under 3W, and supports up to 8GB of memory. The design targets power- and space-constrained applications, including robotics, drones, machine vision, smart cameras and IoT devices.

The multi-engine architecture assigns each sensing or AI workload to the engine best suited to it. SensPro provides programmable processing for sensing and signal-processing tasks, while NeuPro-Nano provides efficient neural processing for always-on inference. Together, the two add flexible sensing and inference capabilities to Mobilint's proprietary AI processing architecture.

REGULUS already enables commercial products, including the MLD-R1 AI USB accelerator, which was previewed at COMPUTEX 2026 and is now shipping. Dongjoo Shin, CEO of Mobilint, said: “REGULUS brings together the specialized processing capabilities needed to run sophisticated AI workloads directly on the device within very tight power constraints.”

Yaron Galitzky, Executive Vice President and General Manager, AI Division at Ceva, said: “REGULUS marks an important milestone for Ceva as the first production deployment of our NeuPro NPU family, demonstrating the progress we are making in bringing our AI technology into commercial silicon.”