According to Jeff Dean on X, Google introduced TPU 8t for large-scale training and inference with a pod size of 9,600 chips delivering about 121 exaflops FP4 per pod, roughly 3X the FP4 performance of Ironwood’s 42.5 exaflops per pod (as reported in Dean’s April 23, 2026 post). According to Jeff Dean, the FP4-focused uplift targets high-throughput inference and frontier model training, signaling lower cost per token and faster time-to-train for multi-trillion parameter workloads. As reported by Jeff Dean, the pod-level scaling implies denser datacenter footprints and higher utilization for Google Cloud customers building LLMs and VLMs, creating business opportunities in model serving, batch inference, and fine-tuning at scale.

According to Jeff Dean on X (formerly Twitter), Google has announced the general availability of its 7th generation TPU, codenamed Ironwood, for Cloud TPU customers. This new release features significant improvements in both performance and efficiency compared to previous generations, enabling faster model training and inference for enterprise AI applications. The Ironwood TPU is expected to accelerate large-scale machine learning workloads, including generative AI and deep learning, providing a substantial competitive advantage for businesses leveraging Google Cloud's AI infrastructure (source: x.com/sundarpichai/status/1986463934543765973).