The Information Machine

To learn more, please check our networking model:

SemiAnalysis Twitter · SemiAnalysis (@SemiAnalysis_) · 2026-06-03

SemiAnalysis launches a detailed AI Networking Model covering hyperscaler cluster configurations, vendor market shares, and pricing for switches, optical modules, and cables across Microsoft, Google, Meta, Amazon, Oracle, X.AI, and neoclouds.

Open original ↗

Appears in

Extraction

Topics: ai-infrastructuredata-center-networkinghyperscaler-analysisoptical-networking

Claims

  • The networking layer of AI and cloud infrastructure is increasingly critical and complex enough to warrant dedicated analytical modeling.
  • Major hyperscalers each use distinct networking configurations depending on accelerator type, spanning scales from H100 to GB300 NVL.
  • SemiAnalysis covers 80+ configuration panels detailing switch, transceiver, cable, AEC, and DAC quantities, pricing, and power consumption per hyperscaler.
  • Optical module market forecasts cover 200G, 400G, 800G, and 1.6T segments through 2026 with vendor-level market share breakdowns.

Key quotes

The SemiAnalysis AI Networking Model is built to decode the increasingly critical networking layer of Al and cloud infrastructure.
This model completes our suite by providing granular visibility into Switches, Transceivers, Cables, AEC/DACs for scale up, scale-out backend, front end, and out of band networks for Al.