Hardware

Google adds TPU 8tx with AMD EPYC Turin hosts and confidential computing, and TPU 8i Max for inference

Google has extended its Tensor Processing Unit line with two new systems, the TPU 8tx and the TPU 8i Max. The TPU 8tx pairs AMD EPYC Turin host processors with confidential computing, while the TPU 8i Max doubles memory bandwidth for inference workloads. Google said the systems respond to external customers asking for accelerators that can reason across large AI model context windows, and to highly regulated customers seeking specialized encryption. The company tied the launch to recent PyTorch work aimed at making TPUs more accessible to developers and Cloud customers.

AI-written summary, checked against the cited sources. How summaries work

Image: TechPowerUp

What we know

Confirmed by several sources

Reported by one source

  • The TPU 8tx pairs AMD EPYC Turin hosts with confidential computing. — ServeTheHome
  • The TPU 8i Max doubles memory bandwidth for inference. — ServeTheHome
  • Google said some highly regulated customers asked for specialized encryption capabilities, and businesses want accelerators that can reason across massive context windows. — TechPowerUp
  • Google said the new systems, together with its recent PyTorch announcements, aim to make TPUs more accessible to every developer and Cloud customer. — TechPowerUp

Open questions

  • It is not yet clear when the new TPU systems will be generally available or how they will be priced. — TechPowerUp

Coverage 2 publishers

  1. TechPowerUpFirst report

    (PR) Google Extends Tensor Processor Unit Series with TPU 8tx and 8i Max

    While we've always co-designed our TPUs in partnership with Google DeepMind, we're now increasingly getting feedback from external organizations on their specific needs. With computing moving from human to agent scale, businesses are asking…

  2. ServeTheHome

    Google TPU 8tx Pairs AMD EPYC Turin Hosts with Confidential Computing while TPU 8i Max Doubles Memory Bandwidth

    The new Google TPU 8tx pairs AMD EPYC Turin hosts with confidential computing while TPU 8i Max doubles memory bandwidth for inference

Earliest report first. Articles stay on their publishers’ sites; each link opens the original.