Google adds dual-chip TPU 8i Max and confidential TPU 8tx
Google has expanded its eighth-generation TPU lineup with TPU 8i Max and TPU 8tx. The company describes them as new system configurations rather than new chip designs. Both are intended for Google Cloud TPU systems, alongside the previously announced TPU 8t and TPU 8i.
TPU 8i Max combines two TPU 8i chips in one package. Google says this doubles compute, on-chip SRAM and HBM bandwidth within the same rack footprint, targeting workloads with very large context windows, including agents handling entire codebases or hours of video. The company has not published absolute specifications, prices or supported regions for the system. Bloomberg, cited by Quartz, reported availability in 2027.
TPU 8tx uses TPU 8t silicon with 5th Gen AMD EPYC host CPUs and the Boardfly network used by TPU 8i. It is designed to protect data and model weights during active processing for regulated organizations and governments. Google claims latency reductions of up to two times compared with TPU 8t, with a near-zero security penalty. The announcement also highlighted TorchTPU, TPU Sync and other software additions for PyTorch, model serving and agent workloads.
Written by Sakhanda Wire AI from reporting by MarkTechPost. Facts, quotes and full context are in the original article below.