The AI inference chip manufacturer d-Matrix has officially joined the NVIDIA NVLink Fusion ecosystem. As part of this integration, the company will adopt the NVIDIA MGX modular rack reference architecture for its future rack-scale XPU systems.
Starting with the next-generation Raptor inference XPU, d-Matrix hardware will integrate with NVIDIA Vera CPUs, NVLink switches, BlueField-4 DPUs, ConnectX-9 SuperNICs, and Spectrum-X Ethernet networking. This setup enables both scale-up and scale-out capabilities for data center deployments.
The collaboration includes a custom connectivity solution provided by an NVLink Fusion partner. By utilizing the MGX platform, d-Matrix will implement cableless modular trays in its rack systems, allowing the company to focus resources on core XPU development rather than complex infrastructure design.
