Google Unveils Ironwood, A TPU for the 'Age of AI Inference' with Enhanced Performance and Efficiency

Google has unveiled Ironwood, its seventh-generation Tensor Processing Unit (TPU), designed for AI inference. Ironwood is the most powerful, scalable, and energy-efficient TPU, engineered to tackle the immense computational demands of next-generation AI, including Large Language Models and Mixture of Experts. Its architecture scales up to 9,216 liquid-cooled chips connected through a cutting-edge Inter-Chip Interconnect network, delivering 42.5 exaflops of compute power and double the performance per watt compared to its predecessor, Trillium.