Google Adds TPU 8Tx and TPU 8I Max to Eighth-Generation Lineup
What happened
On November 9, 2026, Google introduced two new systems, TPU 8tx and TPU 8i Max, to its eighth-generation TPU lineup. TPU 8tx is designed for confidential inference and small-scale training, targeting organizations that handle sensitive data. TPU 8i Max is built to provide higher memory bandwidth and increased rack density for workloads that require processing large amounts of data simultaneously. Both systems use existing eighth-generation TPU silicon but are packaged and configured differently to address specific needs.
Key details
- TPU 8tx uses the same chip as TPU 8t but pairs it with 5th Generation AMD EPYC CPUs and a low-latency Boardfly network.
- TPU 8tx offers up to 2x lower latency than TPU 8t with minimal performance loss.
- TPU 8i Max combines two TPU 8i chips in one package, doubling compute, on-chip SRAM, and high-bandwidth memory compared to TPU 8i.
- TPU 8i Max fits within the same rack footprint as previous models, increasing compute density per square foot and per watt.
- Google reported nearly 100 TPU customers across sectors such as research labs, trading firms, retailers, and universities.
Why it matters
Organizations in regulated industries and those handling proprietary data now have hardware designed to keep information protected even during processing, reducing the risk of data exposure. Data centers and cloud customers can deploy more computing power in the same physical space, which can lower costs and improve efficiency for workloads that require large memory bandwidth. These changes allow Google’s customers to access specialized AI infrastructure without waiting for a new generation of chips, offering faster adaptation to specific technical requirements.