d-Matrix

D-Matrix is developing a platform that enables data centers to handle large-scale generative AI inference with high throughput and low latency. The company’s offering includes the Corsairâ„¢ inference accelerator card, the JetStreamâ„¢ I/O networking accelerator, and the SquadRackâ„¢ rack-scale system, all designed to work together to deliver efficient inference compute. Their technology brings memory and compute closer together, embedding processing inside memory to reduce data movement, improve speed, and energy efficiency. The stack also includes software tools for deployment, model debugging, and performance monitoring, enabling enterprises to scale inference workloads across models and usage scenarios.