AI infrastructure has largely been built around the demands of model training, but inference introduces a different set of requirements for where compute is located, how data moves and how networks connect it all. With workloads extending beyond centralized facilities, network providers must consider how fiber capacity and infrastructure planning can support a wider range of applications. Decisions around network design, infrastructure investment and future demand will play an important role in supporting the next phase of AI adoption. Key topics of discussion include:
- How inference is influencing compute locations and network architecture
- The role of industry partnerships in developing infrastructure that supports varying workloads
- Planning network capacity and infrastructure investments with evolving AI requirements
- How the needs of data centers, hyperscalers and enterprise customers are shaping network priorities
