We clean up after ourselves.

The Juniper QFX5220-128C redefines AI networking density with 128 ports of 100 Gigabit Ethernet in a single 2RU chassis. This ultra-high-density spine switch provides massive scale-out capability for the largest GPU clusters and AI training environments.

With 12.8 Tbps of total throughput, the QFX5220-128C eliminates bottlenecks in distributed machine learning workloads. Deploy in leaf-spine architectures supporting thousands of GPU servers. Each 100G port can break out to 4x25G for leaf switch connectivity or connect directly to GPU servers.

Engineered for hyperscale AI deployments. Supports advanced features like EVPN multihoming, VXLAN overlays, and programmable forwarding. Juniper Apstra provides closed-loop automation for fabric deployment and day-2 operations.

Key Benefits:

• 12.8 Tbps fabric scales to enterprise AI requirements
• 128 ports maximize connectivity in minimal rack space
• Consistent sub-microsecond latency under load
• Automated fabric deployment reduces errors
• Advanced telemetry for AI traffic optimization

Professional services available for fabric design and deployment. Premium support with 4-hour replacement.