d-Matrix builds inference accelerators, networking, and software around digital in-memory computing, reducing data movement to serve generative AI models efficiently at data-center scale.