d-Matrix announces Nvidia partnership to connect Raptor XPUs to MGX systems
d-Matrix says its next-generation Raptor XPUs will feature 3D-stacked DRAM and promises ultra-low-latency AI inference. Customers will be able to run Raptor racks independently or alongside Nvidia GPUs, it says.
TLDR
d-Matrix announced a partnership with Nvidia, saying its next-generation Raptor XPUs will plug directly into Nvidia MGX rack-scale systems through NVLink Fusion. The company says the XPUs will use 3D-stacked DRAM and promises ultra-low-latency inference—generating outputs from trained AI models—for premium token services. It says customers will be able to run a Raptor rack on its own or alongside Nvidia GPUs, matching compute to different phases of a workload.
Combined views
—
2 Sources, first seen ago
Comments