Google launches two TPU variants for secure and long-context AI

The 8tx uses the same silicon as TPU 8t, integrates with AMD EPYC processors, and offers hardware-level data isolation with up to 50% lower latency. TPU 8i Max combines two 8i chips in one package, doubling compute power, on-chip memory, and bandwidth for memory-intensive tasks like large-scale agent workloads and video analysis. Both are part of Google's eighth-gen TPU lineup, with nearly 100 customers including AI labs and financial firms.

Summarised from the original article by an open-source model, then checked automatically: every number and English name must appear in the source. The original article is authoritative.