Global AI Hardware & Accelerators Database · Specs, Benchmarks & Compute Architecture
Location:Home > TPU & Cloud Silicon > Google TPU v6 (Trillium) vs TPU v5p: Optical Circuit Switching, Pod Scaling & Matrix Math Efficiency Specs Sheet

Google TPU v6 (Trillium) vs TPU v5p: Optical Circuit Switching, Pod Scaling & Matrix Math Efficiency

Silicon & Pod Architecture

Google TPU v6, codenamed Trillium, is Google's sixth-generation Tensor Processing Unit tailored exclusively for training and serving massive frontier models like Gemini. Trillium achieves a 4.7x increase in peak compute per chip over TPU v5e while doubling HBM bandwidth and inter-chip interconnect (ICI).

SpecificationTPU v6 (Trillium)TPU v5pTPU v5e
Peak Compute Metric4.7x vs v5e (High FP8/BF16)459 TFLOPS (BF16)197 TFLOPS (BF16)
Memory ArchitectureHigh-Density HBM95 GB HBM2e (2.76 TB/s)16 GB HBM2e (819 GB/s)
Pod ScalabilityUp to tens of thousands8,960 chips in 3D Torus256 chips
Network Topology3D Torus + Optical Circuit Switching (OCS)3D Torus with OCS2D Torus
Power Efficiency67% improvement per WattHigh-efficiency custom PodCost-optimized

Optical Circuit Switching (OCS) Advantage

Unlike traditional spine-leaf Ethernet or InfiniBand networks requiring thousands of expensive electrical switches, Google TPU Pods dynamically reconfigure 3D Torus interconnects via MEMS-based Optical Circuit Switches. This eliminates electrical packet routing overhead and enables fault recovery in seconds by bypassing failing nodes optically.

Copyright © 2026 AI Chips 100 | Global AI Hardware & Accelerators Database All Rights Reserved.  

Sitemap