French semiconductor company developing AI inference infrastructure designed to make AI deployment economically viable. Its inference processor, Jotunn8, is purpose-built to overcome the memory wall and is designed for datacenter-scale inference workloads.