Specifications
Product
Edgecore AIS1600-64O open networking switch
Port Configuration
64 × 1.6T OSFP
Aggregate Bandwidth
102.4 Tbps class
Form Factor
Data centre switch, open networking (OCP-aligned)
Target Network Tier
Native 1.6T or twin-port switch tier feeding 800G NICs via 2×800G breakout
Optics Support
1.6T OSFP modules; 800G OSFP / OSFP-RHS / QSFP-DD breakout
Announcement
February 2026
Peer Platforms
Broadcom Tomahawk 6 systems, Arista 7060XE7 family, Marvell Teralynx T100
Ecosystem
Open networking software stack, hyperscale AI cluster deployments
Use Case
AI training and inference fabric spine layer
Overview
The Edgecore AIS1600-64O is a 64-port 1.6T OSFP open networking switch announced in February 2026 for hyperscale AI clusters. It belongs to the first generation of 102.4 Tbps-class platforms built to carry native 1.6T optical modules.
Deployment practice in 2026 pairs a native 1.6T or twin-port switch tier with 800G-class GPU NICs using 2x800G breakout, so the 1.6T spine can be introduced while host ports remain at 800G. Edgecore positions the AIS1600-64O as an open-networking alternative to branded platforms in the same class.
Key Benefits
102.4 Tbps-class switching in a single 1U-class platform; 64 OSFP cages accept 1.6T modules or 2x800G breakout for staged migration; open networking software avoids single-vendor lock-in; built for RoCEv2 AI fabrics with the lossless behaviour large GPU clusters require.
Applications
Hyperscale AI training and inference fabrics, GPU cluster spine and leaf layers, research HPC interconnect and open-networking data centre builds.
Request a Quote — EDGECORE AIS1600-64O — 64×1.6T OSFP OPEN AI DATA CENTRE SWITCH
QS Compute — global B2B supply of AI computing hardware, edge AI systems and accelerators. Volume pricing, 15-day sample lead time.
Get Your Quote →