NVIDIA is reportedly planning to invest in d-Matrix, an AI inference ASIC startup, following its de facto acquisition of Groq. d-Matrix had previously collaborated with NVIDIA on technology, with its next-gen inference XPU supporting NVLink Fusion and the MGX rack architecture.
AI inference chip maker d-Matrix announced on September 10 local time that it has joined NVIDIA's NVLink Fusion ecosystem. Its rack-scale XPU systems will adopt the NVIDIA MGX reference architecture, and next-generation inference XPUs starting with Raptor will integrate with NVIDIA Vera CPUs and NVLink switches for scale-up and scale-out.