Large Language Models (LLMs) are transforming AI, but they face critical challenges in grounding, memory, and reasoning. This session explores how graphs (from knowledge graphs and ontologies to graph databases and graph neural networks)address these gaps and are essential for the future of LLMs. We begin with core concepts and academic foundations, then move into practical applications in search, recommendations, and reasoning. Global case studies and live examples from skilling, healthcare, and legal domains will be shared. Participants will also learn about key commercial and open source tools like Neo4j and Janus, the role of ontologies, and how LLMs are making ontology creation faster and easier.