What's new for AI on GKE: Training, serving, and agents

The requirements for AI infrastructure are shifting rapidly, moving from frontier model serving to complex reinforcement learning and agentic workflows. Google Kubernetes Engine (GKE) has evolved to meet the moment. Join us for a high-speed tour of most critical GKE innovations released in the past year for AI training, serving, reinforcement learning, and agentic workloads. Explore new capabilities designed to make your AI workloads more reliable, scalable, and cost-effective. And take a peek at how Thinking Machines Lab and Baseten are using GKE for their AI workloads. Watch more Google Cloud Next 2026 Sessions → https://goo.gle/gcloud-next-2026 Subscribe to Google Cloud Tech → https://goo.gle/GoogleCloudTech Speakers: Nathan Beach, Colin McGrath, Myle Ott BRK2-118 #GoogleCloudNext