Google Takes Aim at Nvidia With New Tensor Chips to Power AI Boom

In brief

• Google introduced eighth-generation Tensor Processing Units with two architectures: TPU 8t for training and TPU 8i for inference.
• TPU 8t delivers nearly 3x the compute performance per pod over previous generations, scaling to 121 ExaFlops.
• TPU 8i features 3x more on-chip memory to…