d-Matrix Adopts NVIDIA NVLink Fusion for Rack-Scale XPU Deployment
NVIDIA, Thursday, September 10th, 2026
d-Matrix integrates its Raptor XPUs via NVLink Fusion, slotting specialized silicon into existing NVIDIA rack architecture.
Chipmaker d-Matrix is integrating its Raptor XPUs using NVIDIA NVLink Fusion, letting the specialized processors deploy inside existing NVIDIA rack architecture, networking and supply chain rather than requiring separate infrastructure.
The strategic pattern is significant: NVIDIA is positioning its interconnect and rack design as the substrate that alternative accelerators plug into, which keeps it central even where its own GPUs are displaced.
For infrastructure buyers, it means specialized inference silicon may become deployable without a parallel operational stack.