Neuronova is a fabless semiconductor company based in Milan developing fully-analog neuromorphic processors for always-on audio intelligence and ultra-low-power edge computing.
Its H1 processor runs voice activity detection, wake-word detection, keyword spotting, acoustic scene classification and gate-assisted denoising at sub-microwatt end-to-end power - orders of magnitude below conventional digital approaches.
H1 is fully programmable, so the same always-on co-processor can be retargeted across use cases as product requirements evolve. For earbuds, hearables, smart glasses and other wearables this is not only a battery question: within the same energy budget, designers can either extend runtime or enable features that were previously out of reach.
H1 is built on European silicon at a European foundry. The company offers validated silicon, an evaluation kit and a PyTorch-compatible SDK for on-device AI and TinyML development
https://www.neuronovatech.com/
Team contacts
