Memory-centric AI inference accelerator card
Corsair is d-Matrix's inference-optimized accelerator that brings high-bandwidth memory and compute together on the same fabric. Public materials describe an inference-optimized compute engine, chip-to-chip connectivity, and a standard PCIe form factor that can run as a standalone inference solution or alongside GPUs.