Google presenteert nieuwe TPU-generatie voor AI-agenten en modeltraining

Google has unveiled its latest generation of Tensor Processing Units (TPUs), featuring two specialized chips designed to accelerate model training and agent workflows that require continuous, multi-step reasoning and distributed action loops across multiple models. The new TPUs offer improved performance, memory capacity, and energy efficiency compared to previous generations. These chips are specifically optimized for training state-of-the-art models and supporting complex agent-based applications.