Why Enterprise Data Centers Still Matter

Hyperscale AI campuses command the headlines, but the next major wave of AI adoption may play out across enterprise data centers measured in megawatts rather than hundreds of megawatts. In this episode of the Data Center Frontier Show, DCF Editor in Chief Matt Vincent sits down with Kirk Killian, President of Partners National Mission Critical Facilities, to examine how Fortune 1000 and Global 2000 organizations are preparing for AI—and why their infrastructure priorities differ sharply from those of hyperscalers. Killian argues that while enterprises are comfortable outsourcing AI training, the rise of AI inference could drive sensitive workloads back toward on-premises environments and private colocation deployments, where latency, security, compliance, and operational control become paramount. He also explains why enterprise customers continue to prioritize reliability and flexibility over sheer scale, how cabinet densities are evolving, why liquid cooling optionality matters even when it's not immediately needed, and what developers can do to better serve this often-overlooked market. The conversation also explores the future of hybrid cloud, the economics of AI infrastructure, emerging enterprise site selection trends, and why “cloud plus controlled” may become the dominant architecture for enterprise AI. For anyone focused on the next phase of AI infrastructure—not just the largest campuses, but the environments where AI will be embedded into everyday business operations—this discussion offers an important and frequently overlooked perspective.