<ul><li>Google Cloud is reshaping the future of AI infrastructure to accommodate more complex models and increasing enterprise demands.</li><li>Mark Lohmeyer, VP and GM of AI computing infrastructure at Google Cloud, discusses the company's open-source AI efforts and the importance of performance and cost efficiency.</li><li>The emergence of reasoning models and intense computational loads in the inference era necessitates powerful, adaptable, and cost-efficient infrastructure to meet the demands of multi-step decision-making models.</li><li>Google is extending VLLM support to TPUs, enhancing price-performance potential for customers and emphasizing the importance of open-source technologies like JAX for AI model training and serving.</li></ul>

Google Cloud retools AI infrastructure for the inference era

Discover more