d-Matrix Joins NVIDIA NVLink Fusion Ecosystem

d-Matrix, an AI inference chip maker, announced on September 10 that it has joined NVIDIA's NVLink Fusion ecosystem. The company's rack-scale XPU systems will be built on NVIDIA's MGX modular reference architecture.

Starting with the next-generation inference XPU Raptor, d-Matrix XPUs will integrate with NVIDIA Vera CPUs, NVLink switches, BlueField-4 DPUs, ConnectX-9 SuperNICs, and Spectrum-X Ethernet networking, enabling both scale-up and scale-out. Another NVLink Fusion partner will provide a custom connectivity solution for d-Matrix XPUs, and the MGX-based racks will use modular cable-free trays.

Joining the NVLink Fusion ecosystem allows d-Matrix to adopt rack-level solutions with larger scaling domains without complex selection and design work, allowing the company to focus more on core AI XPU development.