Google adds TPU 8tx and TPU 8i Max systems

Google expanded its eighth-generation TPU line on November 9 with TPU 8tx, a confidential-computing system that keeps data protected while it is being processed, and TPU 8i Max, which pairs two TPU 8i chips in one package for double the compute and memory bandwidth.

Why it matters: The two systems take Google's TPU portfolio to four launches in one year and sharpen its pitch to regulated enterprises and agent workloads, the segments where Nvidia's grip on AI accelerators is most contested.

Data as of Google announcement of November 9, 2026, read in full in this run; performance figures are Google's own claims, not independently benchmarked. Yonhap report of November 9, 2026, in Korean, read in full via the estate browser route, for the Google Cloud Next Tokyo venue. Google gave no availability date or price for either system. Developing item.

Google added two systems to its eighth-generation Tensor Processing Unit line on November 9. TPU 8tx is built for secure and confidential inference: it uses the same silicon as the TPU 8t training system with fifth-generation AMD EPYC host processors, and keeps sensitive data and model weights protected in hardware while they are in active use, a step aimed at regulated enterprises and governments. TPU 8i Max combines two TPU 8i chips on a single package, delivering double the compute, double the on-chip SRAM and double the high-bandwidth memory capacity of the 8i, for workloads that run across very large context windows.

The additions make 2026 the first year Google has introduced four TPU systems, following the 8t and 8i in April, and Google said its TPU business now serves nearly 100 customers. The announcement, made at Google Cloud Next in Tokyo according to Yonhap, carries no availability date or pricing for either new system.

Google adds TPU 8tx and TPU 8i Max systems

Sources