Google announced two additions to its eighth-generation Tensor Processing Unit hardware lineup, the TPU 8tx and the TPU 8i Max.
In the announcement, released Nov. 9, 2026, the company said the systems are designed to manage complex artificial intelligence tasks by supplying specialized computing for security and reasoning. The additions bring the company’s total number of eighth-generation processing systems to four.
The TPU 8tx operates alongside fifth-generation AMD EPYC processors. Google stated the system builds security features directly into the hardware layer to isolate sensitive data while it is in active use. The company said this allows organizations to adjust machine learning models using private data without moving it between separate computer clusters.
For the TPU 8i Max, Google combined two TPU 8i chips onto a single package. The company stated this design is intended to prevent memory bottlenecks when programs process large volumes of data. Google said the dual-chip configuration fits twice the silicon into existing racks, allowing data centers to increase computing power per square foot.
The release also detailed software updates, including an open-source layer called TPU Sync to manage memory across systems. The company said its Cloud Pathways and dynamic slicing tools are now generally available on the Google Kubernetes Engine to coordinate tasks across multiple hardware units.