The state of play of data center development

In this episode of Catalyst, Shayle Kann interviews Chris Sharp, CTO of Digital Realty, about the current state and future of data centers, particularly in the United States. They discuss the geographic distribution of data centers, the impact of AI workloads (training vs. inference) on data center demand, and the challenges of power constraints and grid interconnection. The conversation covers the importance of availability zones, the increasing sophistication of AI models, and the trend towards regional data center growth. They also delve into the practical aspects of data center development, including timelines, equipment bottlenecks, and the role of bridge power solutions. Chris shares his insights on the demand for large-scale data centers, the need for densification, and the importance of partnerships with utility operators. The episode concludes with Chris's excitement about advancements in data center technology, particularly in efficiency and hardware design, and the potential of AI to benefit society.