AI Briefing
KOSign in

Google adds TPU 8tx and 8i Max to 8th-gen portfolio, shipping four new TPU systems in one year

·2026.11.09 22:00

Key point

TPU 8tx offers up to 2x lower latency for confidential compute, while TPU 8i Max doubles compute and HBM bandwidth to resolve memory bottlenecks.

1 / 9

Details

  • Google announced the addition of TPU 8tx and TPU 8i Max to its 8th-generation TPU portfolio, marking the release of four new TPU systems (8t, 8i, 8tx, 8i Max) within one year.
  • TPU 8tx is optimized for confidential inference in regulated industries, providing hardware-level security and up to 2x lower latency compared to TPU 8t with near-zero performance penalty.
  • TPU 8i Max integrates two TPU 8i chips into a single package to address the memory wall, delivering 2x compute, 2x SRAM, and 2x HBM bandwidth compared to TPU 8i.
  • New software tools include open-sourced TorchTPU for native PyTorch execution, TPU Sync for reducing KV transfer latency and TTFT by over 90%, and GKE dynamic slicing using Optical Circuit Switching to cut job start times by up to 5x.
  • Google DeepMind's AlphaChip and Gemini are used to accelerate chip design, reducing layout generation time from weeks to hours.
googleai-infrastructuretputorchtpuoptical-circuit-switchingtpu-8txtpu-8i-maxconfidential-compute
Read originalShare on Threads

This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.

Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.

Related

RedditOpenAI Reportedly Plans to Buy Google TPUs to Reduce Nvidia RelianceGoogle AI BlogGoogle Launches Project Suncatcher Prototype to Test TPU Survival in SpaceGeekNewsGoogle acquires Crusoe's Wyoming data center siteGoogle Cloud AIGKE dynamic slicing for TPUs is now generally available, reducing startup times by up to 5xGoogle AI BlogGoogle receives first AI response from Project Suncatcher prototype satelliteHacker NewsNvidia Used Circular Financing to Keep GPU Prices High, WSJ Report Claims