Google Adds TPU 8tx and 8i Max to Eighth-Gen AI Chip Lineup
Google introduced TPU 8tx and TPU 8i Max on Monday, expanding its eighth-generation AI chip lineup to four systems in one year.
Alphabet's Google on Monday introduced two new variants of its eighth-generation Tensor Processing Units, TPU 8tx and TPU 8i Max, expanding the lineup to four systems in a single year for the first time.
TPU 8tx is built for secure and confidential inference. It is aimed at government agencies and companies in regulated industries that need to keep sensitive data protected while it is being processed. The chip pairs Google's TPU silicon with 5th Generation AMD EPYC processors and encrypts data at the hardware level, keeping data and model weights protected even while in active use. Google said it delivers up to two times lower latency than TPU 8t, with near-zero performance penalty, and is also suitable for small-scale training.
TPU 8i Max addresses what Google calls the memory wall. It combines two TPU 8i chips into a single package, delivering double the compute, double the SRAM and double the HBM bandwidth of TPU 8i. The company said it is suited to applications that process very large amounts of information at once, such as coding assistants, mathematical reasoning and drug discovery, and offers the highest compute density per rack.
The two new systems build on TPU 8t and TPU 8i, introduced in April. Google said those chips offer up to 2.7 times better training performance and 80 percent better inference price performance than the previous generation. Together the four systems cover training, real-time reasoning, confidential computing and large-context workloads.
Google's TPUs were long used mainly inside the company for products such as Search, YouTube and Gemini. The company has been making them available through Google Cloud to attract enterprise AI customers. TPU sales made a meaningful contribution to results in the September quarter, according to Alphabet, with Google Cloud revenue more than doubling.
Alphabet shares were up more than 3 percent in early trading on Monday, Reuters reported. In delayed trading the stock was up 1.18 percent at $288.15.
Bloomberg News first reported the development; Google's chief technologist for AI infrastructure, Amin Vahdat, told Bloomberg the 8tx and 8i Max are reconfigured versions of existing chips designed for workloads customers have been asking for.