NVIDIA Spectrum-4 SN5600 128×400GbE GPU Top-of-Rack Switch
One 2U platform splits to 128 ports of 400GbE per rack.
We help you choose, configure, and deliver the right system — no obligation.
Configuration at a Glance
Tailored per engagement. Full technical overview below.
Overview
Configured as a high-radix top-of-rack switch, the Spectrum-4 SN5600 breaks its 64 OSFP cages out to 128 ports of 400GbE for dense GPU server connectivity. It is specified, split-configured, tested, and warranty-backed, sourced through NVIDIA-authorized channels for predictable rack-level deployment.
Who This Solution Is For
Business Benefits
Highest 400GbE radix
Up to 128 ports of 400GbE in 2U serve a full rack of GPU servers from a single switch.
Lower switch count
Doubling port density at the leaf reduces the number of ToR switches, optics, and management points.
Configured per rack
Breakout profiles and cabling plans are set and validated before shipment for fast, repeatable rack turn-up.
Typical Business Use Cases
Top-of-rack for 400GbE GPU servers
Dense leaf layer in AI clusters
Rack-scale RoCEv2 connectivity
Mixed 400/800GbE fabric edges
Industry Applications
Technical Overview
The same NVIDIA Spectrum-4 ASIC drives 51.2 Tb/s while its 64 OSFP ports split into 128× 400GbE interfaces with full wire-speed, low-latency cut-through forwarding. RoCEv2 congestion control and adaptive routing make it a high-density leaf for Spectrum-X Ethernet fabrics.
| Switch ASIC | NVIDIA Spectrum-4 |
| Port Configuration | 128× 400GbE (64 OSFP split) |
| Switching Capacity | 51.2 Tb/s |
| Packet Rate | 33.3 billion packets/sec |
| Fabric Features | RoCEv2, adaptive routing, cut-through |
| Form Factor | 2U, OSFP with breakout |
| Network OS | Cumulus Linux / SONiC (open) |
| Management | 1× 25GbE SFP28, dedicated mgmt + console |
| Warranty | Manufacturer-backed, authorized channel |
Specifications are indicative and configured to each engagement. Request a quote for a configuration tailored to your requirements.
Warranty, Support & Fulfillment
Every system ships from an authorized channel, configured and tested, with the documentation enterprise buyers need — backed by warranty and a dedicated account team.
Enterprise Warranty
Full manufacturer warranty with optional on-site, next-business-day support and extended coverage.
Authorized Channel
Sourced through Tier-1 distribution and OEM partners — never grey market. Asset & warranty records included.
Lead Time & Deployment
48-hour quotes, then configured, burn-in tested, and delivered on a committed schedule.
Nationwide Fulfillment
Coordinated logistics, rack-and-stack, and delivery wherever your infrastructure lives.
Frequently Asked Questions
Why split 800GbE down to 400GbE at the rack?
Most current GPU servers connect at 400GbE, so splitting maximizes server attach per switch and matches NIC speeds while preserving an 800GbE-capable platform for future upgrades. We pre-set the breakout to your server fleet.
How many GPU servers fit per switch?
At 128 ports of 400GbE a single SN5600 can attach a full rack of dual- or quad-port GPU servers; exact counts depend on per-server NIC count, which we map during sizing.
Is this the same hardware as the 800GbE model?
Yes — identical Spectrum-4 hardware, configured for high-radix 400GbE leaf duty instead of an 800GbE spine. The breakout and OS profile are validated before delivery.
Hardware Assistance
Configure the NVIDIA Spectrum-4 SN5600 128×400GbE GPU Top-of-Rack Switch with Nexus Compute
Tell us your requirements and a hardware specialist will help you specify, configure, and quote the right system — typically within two business days. No obligation.