Mark Lohmeyer, Vice President and General Manager of Compute and Artificial Intelligence Infrastructure at Google Cloud, joins Rob Strechay of theCUBE Research and Rebecca Knight of SiliconANGLE Media during Red Hat Summit 2025 to discuss Google Cloud’s latest collaborations and advancements in artificial intelligence infrastructure. In this segment, Lohmeyer explores Google Cloud's focus for 2025, which centers on inference workloads and the evolution of their infrastructure to support increasingly complex AI models. They emphasize Google’s strong presence in the open-source community, referencing technologies such as JAX and Kubernetes, and discuss the emerging synergies with Red Hat to address AI challenges, particularly in reasoning models. Key points from the discussion include Google’s initiatives in low-cost high-performance AI solutions using tensor processing units and graphics processing units, as well as the open-source project LLMD, which offers flexibility across various computing needs. Lohmeyer states that Google’s strategic collaborations aim to enhance customer outcomes, focusing on a consistent platform that provides both choice and operational consistency across vendors and infrastructures.