d-Matrix brings Raptor chips to Nvidia racks as falling token prices pressure GPU economics

DIGITIMESSeptember 10, 2026en

AI chip startup d-Matrix is adopting Nvidia's NVLink Fusion technology to integrate its next-generation Raptor inference processors directly into Nvidia MGX racks, giving the company a deployment path for the heterogeneous computing architecture it says will become increasingly important as the economics of GPU-based inference come under pressure.

This is a short summary published by AI Global Wire. The full article is owned and hosted by DIGITIMES — open it there to read it in full.