In this keynote session focused on enterprise generative AI infrastructure, our speaker will explore strategies for optimizing efficiency to align with the demand. The discussion will unravel key insights into designing and implementing infrastructure tailored to generative AI, ensuring a harmonious balance between computational resources and the evolving needs of enterprise applications. Beneath this vast AI landscape lies a pivotal force: AI inference. It operates like the enigmatic wizard behind the curtain, tirelessly deploying and running trained AI models and applications within the world’s data centers day in and day out. But here is the kicker: while AI is booming, AI inference remains a blind spot. The costs, complexities, and computational demands of managing AI workloads are skyrocketing. And that IT/AI operational expense will only grow – as much as 8x the cost of AI training - as demand for power-hungry generative AI, large language models and consumer desire for AI-enabled services keep growing. Enter Moshe Tanach, with a bold alternative. He advocates for a paradigm shift from CPU-centric to AI-centric system architecture. This move aims to liberate data centers from performance inefficiencies, unsustainable operating costs, and high energy consumption through the implementation of a new AI system architecture. Join Moshe in championing Affordable Intelligence through AI-centric data centers – a turning point on the AI compute continuum.