Google just expanded its AI chip lineup — and it is taking direct aim at Nvidia.

Google has introduced two new systems in its eighth-generation TPU portfolio: TPU 8tx and TPU 8i Max. Announced on November 9, they complete a four-system lineup alongside TPU 8t and TPU 8i from April — the first time Google has ever introduced four TPU systems in a single year.

TPU 8tx is built for secure, confidential AI. It runs on the same silicon as TPU 8t, paired with 5th-generation AMD EPYC CPU hosts, and builds security into the hardware layer, keeping sensitive data and model weights protected even while in active use. Google says it delivers up to 2x lower latency than TPU 8t, and it is aimed squarely at governments and regulated industries running AI agents on sensitive data.

TPU 8i Max attacks the memory wall. It combines two TPU 8i chips onto a single package, delivering double the compute, double the SRAM and double the HBM bandwidth of TPU 8i in the same rack footprint — built for massive context windows, and well suited to coding assistants, mathematical reasoning and drug-discovery workloads.

Why it matters: AI is moving from human scale to agent scale, and Google is racing to give enterprises a purpose-built chip for every stage. Google Cloud revenue more than doubled to $30.71 billion in the September quarter, and Alphabet shares rose more than 3% on the news. Google has not announced pricing or availability; Bloomberg reports the new systems are expected in 2027.

Source: Google Blog, Reuters