AI Networking — Products
Updated 6/19/2026
Engine-synthesised product landscape for AI Networking, ranked by trend signal across hiring, capital, orders, and discussion axes.
Last refresh: 2026-06-18.
Lossless Ethernet / RoCEv2 AI Fabric (Ultra Ethernet) — fabric architecture / protocol
Trend: 🔥 hot
Opportunity: GPU training/inference fabrics demand RDMA-class semantics on Ethernet; UEC 1.0 targets ~1M endpoints and ESUN extends Ethernet into scale-up. Community discussion is explicitly tracking the InfiniBand → UE transition.
Architectural inflection where three incumbents are committing PM and silicon talent. ESUN + UEC give it standards backing. Open competitive question vs NVIDIA Spectrum-X / MRC and reintroduced Omni-Path.
Companies committing: Juniper Networks, Arista Networks, Cisco.
800G / 1.6T AI Ethernet Switching Platform (51.2T–102.4T class) — data center switch
Trend: ↗ rising
Opportunity: Hyperscaler GPU clusters scaling past 100K endpoints need merchant-silicon, lossless 800G/1.6T Ethernet at radix levels InfiniBand cannot match cost-effectively (fact e070c95f Tomahawk 6 102.4T; fact 9b6493fe Arista $750M FY25 AI back-end).
Clear breakout product: Arista, Cisco, and Celestica (around Broadcom Tomahawk 6) are all converging on 51.2T → 102.4T Ethernet as the de-facto AI back-end fabric, with hyperscaler design wins already monetizing. Remaining gap is congestion control / lossless behavior parity with InfiniBand at scale.
Companies committing: Arista Networks, Cisco, Celestica, Semtech.
Active Electrical Cable (AEC) / Active Copper Cable for AI Racks — high-speed copper interconnect
Trend: ↗ rising
Opportunity: Inside-rack GPU-to-switch links need lower power/cost than optics at sub-3m reach; AWS Trainium and Nvidia rack-scale designs are productizing this exact slot (facts ba213d85 1.6Tbps AEC; cb28d21e Semtech >100% YoY datacenter).
Hot, concentrated: Credo (AWS anchor, ~80% revenue concentration per fact 169a5e0e) and Semtech CopperEdge are the two pure-plays. AEC is now an established AI rack-scale product, not a thesis — the open question is multi-customer diversification before optics-at-1m become viable.
Companies committing: Credo Technology, Semtech.
Optical DSP & SerDes Retimer Silicon for 800G/1.6T — mixed-signal silicon
Trend: ↗ rising
Opportunity: 200G-per-lane PAM4 is the bottleneck for 1.6T pluggables; demand for SerDes/retimer IP is structural (fact 23be5595 Credo SerDes hiring; fact ba213d85 200G/lane DSP).
Quiet but critical layer under the 1.6T transition. Credo monetizes through merchant DSP, Celestica integrates at the module/system level. Risk: hyperscaler in-house SerDes (AWS, Google) could disintermediate.
Companies committing: Credo Technology, Celestica, Semtech.
NVIDIA Spectrum-X / InfiniBand AI Fabric — InfiniBand + Ethernet AI fabric (NIC + switch + protocol)
Trend: ↗ rising
Opportunity: Vertically-integrated alternative to merchant Ethernet at gigascale; community is actively comparing MRC vs UE (demand_query 21f9b813).
Incumbent leader — biggest AI cluster deployments to date sit on this stack. Pressure comes from UEC-based Ethernet ecosystem; NVIDIA is countering with custom MRC transport.
Companies committing: NVIDIA.
Cisco Silicon One AI/ML Switching ASIC — switching ASIC
Trend: → steady
Opportunity: Cisco's bet to compete head-on with Broadcom Tomahawk in the AI back-end; fact d4537830 confirms FY25 AI orders ~$2B (2× target) and fact 5bb3849e confirms ongoing silicon hiring.
Hot for Cisco specifically — one of the few non-merchant-silicon answers to Tomahawk 6. Risk: hyperscalers prefer multi-source merchant silicon, so Silicon One must keep winning on per-port economics.
Companies committing: Cisco.
Liquid-Cooled AI Server / Switch Platform — rack-level system
Trend: → steady
Opportunity: Thermal density at 102.4T switching + 1kW+ GPU servers makes liquid cooling structurally non-optional; Celestica's revenue ramp confirms hyperscaler pull (fact 3dbe13aa $12.2B FY25).
Hot ODM/EMS-side product. Celestica is the dominant signal in this dataset; broader competition (Foxconn, Quanta, Wiwynn) is implied but not captured here.
Companies committing: Celestica.
PCIe Gen5/Gen6 Retimer Silicon for AI Servers — PCIe retimer / connectivity silicon
Trend: → steady
Opportunity: PCIe-lane fanout to GPUs/CXL memory drives a structural retimer attach inside every AI server; community is also probing the adjacent CXL-fabric layer.
Single-vendor dominance: Astera with all top-4 US hyperscalers. The risk is hyperscaler in-house SerDes; the upside is CXL fabric becoming the next attach.
Companies committing: Astera Labs.
Co-Packaged Optics (CPO) for AI Cluster Switches — optical engine / package
Trend: · weak signal
Opportunity: Power and reach limits of pluggable optics at 1.6T/3.2T push the industry toward CPO; community is actively probing the thesis (demand_query ff010e80; fact 80237ba4 Celestica CPO posting 2026-05-06).
Earlier-stage than AEC/DSP. Celestica is the only company in this dataset committing engineering to CPO at AI-cluster scale. Real opportunity if 1.6T → 3.2T transition forces CPO; risk of being premature if pluggable LPO/LRO keeps stretching.
Companies committing: Celestica.
AIOps / Self-Driving Network Platform (Mist AI) — AI-driven network management SaaS
Trend: · weak signal
Opportunity: Enterprise network ops still chronically under-automated; AI-driven anomaly detection on telemetry is a real product gap Juniper continues to invest in.
Adjacent to (not central to) the AI-fabric story but still strategically relevant — Juniper's AI value capture sits in network management software, not silicon. Single-vendor signal in this dataset.
Companies committing: Juniper Networks.