Inference Infrastructure & ServingGoogle Cloud (Cloud Next)Google Cloud

Google Cloud previews its 8th-generation TPUs at Cloud Next

Previewed its 8th-generation TPUs — split into TPU 8t (training) and TPU 8i (inference)

Why it matters

Split training and inference silicon for the first time; a signal that the era of general-purpose accelerators is ending

Source — Primary sources first; secondary reports are labelled “Press”

Pressstoragereview.comhttps://www.storagereview.com/news/google-announces-tpu-8t-sunfish-and-tpu-8i-zebrafish

Added in Week 35, 2026 · See that week

Same week

DeepSeek-V4PaperDeepSeek releases V4-Pro (1.6T/49B) and V4-Flash (284B/13B), MIT, 1M contextOfficialCohere announces acquisition of and merger with Germany's Aleph AlphaPressGPT-5.5Official

Nearby on the same axis

Google Cloud makes TPU7x generally available (GA)OfficialvLLM introduces Model Runner V2OfficialHugging Face archives the text-generation-inference repositoryOfficialNVIDIA's Vera Rubin NVL72 enters volume productionOfficial