d-Matrix Adopts NVIDIA NVLink Fusion for Rack-Scale XPU Deployment
nvidia ·
AI inference chipmaker d-Matrix announced it will adopt NVIDIA NVLink Fusion to connect its next-generation Raptor XPUs to NVIDIA's AI infrastructure platform, becoming part of a growing ecosystem of partners using NVIDIA's rack-scale architecture. By connecting Raptor to NVIDIA NVLink scale-up and Spectrum-X scale-out networking along with the NVIDIA MGX rack architecture, d-Matrix gains an accelerated, lower-risk path from custom silicon to large-scale deployment. d-Matrix cofounder and CEO Sid Sheth said inference demand is soaring while capital, time and energy remain limited, and that NVLink Fusion and MGX let the company integrate its Raptor XPUs into a broadly deployed, liquid-cooled architecture for ultralow-latency inference. According to the article, NVLink Fusion extends NVIDIA's horizontally open platform to XPUs and CPUs, letting silicon companies focus on processor innovation while relying on NVIDIA infrastructure to deploy at AI factory scale. d-Matrix plans to connect its XPUs in a single high-bandwidth, low-latency scale-up domain, and its racks can work alongside NVIDIA GPU-based systems such as Vera Rubin NVL72 for disaggregated inference. The company also plans to integrate NVIDIA Vera CPUs, ConnectX-9 SuperNICs, BlueField-4 DPUs and Spectrum-X Ethernet networking. The NVIDIA full-stack AI factory platform includes Vera Rubin NVL72, Groq 3 LPX, the Vera CPU rack, Vera BlueField-4 STX storage and Spectrum-6 SPX Ethernet networking, and is designed to be fungible across AI workloads. NVLink Fusion is positioned as giving customers flexibility to match compute to each workload within a common AI factory platform.
nvidia