Overview

Credo OmniConnect uses 112G VSR SerDes and AXI framing to deliver scalable, low-latency AI interconnect, solving memory bottlenecks for advanced inference workloads.

Credo OmniConnect is a next-generation interconnect architecture designed to break through the memory bottlenecks that limit AI inference scalability. It combines ultra-efficient 112G VSR SerDes with a lightweight AXI framer to deliver high-bandwidth, low-latency, and power-optimized connectivity between compute engines and memory—whether on-die, off-substrate, or across chiplets.