Powering AI with 224G LR SerDes and Co-Packaged Copper
For next‑generation high‑bandwidth data centers scaling toward 1.6 Tbps, 200G/lane I/O architectures, and high‑density switching for AI/ML training clusters, this 224G SerDes is optimized for co‑packaged copper (CPC) inside next‑generation switches, within the rack, and across the pod. It delivers extremely high throughput while meeting strict power and latency requirements.
These co‑packaged architectures extend copper reach inside the rack and provide a scalable path toward multi‑terabit switching systems and future AI cluster fabrics.
This demo showcases the performance of a Marvell 224G long‑reach SerDes built on the Marvell 3nm silicon platform, using co‑packaged copper to drive traffic across a 2.5‑meter copper cable, demonstrating its readiness for next‑gen data center deployments.
