Table of Contents
Digital Realty’s CEO Andrew Power told investors that the broader digital infrastructure market is evolving rapidly as AI adoption increases demand for capacity while also raising the value of connected environments, strategic locations, and global reach
In sum – what to know
AI demand broadens – Digital Realty says AI demand is extending beyond large-scale training into inference, enterprise adoption, connectivity-rich environments, and new geographies.
Leasing hits records – Third-quarter bookings reached $855 million, while the company’s backlog climbed to a record $2.5 billion.
Inference goes global – AI inference-related deployments are emerging across more than 15 markets, increasing demand for distributed, connected infrastructure.
Digital Realty reported record leasing in the third quarter as artificial intelligence continued to drive demand for data center capacity while increasing the importance of connectivity and distributed infrastructure.
Total third-quarter bookings reached approximately $855 million, according to CEO Andrew Power. The company said bookings across several categories reached records, while its total backlog increased to a new high of $2.5 billion.
Power said the broader digital infrastructure market is evolving rapidly as AI adoption increases demand for capacity while also raising the value of connected environments, strategic locations, and global reach.
Digital Realty said the demand profile is also expanding beyond the large-scale training deployments that have dominated much of the discussion around AI infrastructure. “We continue to see the benefits of artificial intelligence extending well beyond large-scale training deployments and into enterprise adoption, inference workloads, connectivity-rich environments and new geographies,” Power said.
That broader demand was reflected in the company’s colocation and connectivity business. Bookings in the 0-1 megawatt plus interconnection category reached more than $150 million at Digital Realty’s share, up 80% year over year and 40% from the second quarter, marking the fourth consecutive quarterly record in the category.
Power said connectivity is becoming increasingly important as digital infrastructure architectures become more distributed and data-intensive. “While AI is often discussed in terms of large-scale compute deployments, the reality is that AI workloads increasingly depend on a broader digital infrastructure architecture that connects compute and storage with cloud platforms, networks, customers and inference locations,” he said.
The company’s greater-than-1-megawatt business also reached a quarterly record, with total bookings of $689 million, or $474 million at Digital Realty’s share. Demand was broad-based across the company’s footprint, with strong activity in APAC as well as the U.S.
AI inference is becoming an increasingly visible component of that demand.
During the quarter, Digital Realty said a major AI platform was deploying inference and low-latency edge hubs in 15 different metros using the company’s interconnection capabilities and global footprint.
The company also cited a global medical device and services provider deploying distributed AI data hubs in four locations across its three global regions, while a global financial services company was deploying across three major financial centers to meet low-latency requirements.
CFO Matt Mercier said AI inference-related deployments during the quarter spanned more than 15 separate markets, indicating that adoption is moving beyond large training deployments into a broader range of enterprise and inference-oriented use cases.
He identified two areas gaining momentum: agentic AI, where general-purpose agents are beginning to roll out, and sovereign AI, where governments and regulated enterprises require in-region capacity and connectivity to keep models and data within their borders.
The company’s record leasing activity is also feeding a substantial development pipeline. Digital Realty invested $1.5 billion in development capital expenditures during the quarter, while its development pipeline expanded to 1.5 gigawatts under construction at a total expected cost of $22 billion.
Digital Realty also reported that it had delivered 50 megawatts of new IT capacity during the quarter, 95% of which was pre-leased, and commenced development of another 112 megawatts across projects in APAC, EMEA, and the Americas.
Power said the company’s global platform, which spans six continents and includes colocation and connectivity hubs, hyperscale data centers, and strategic private capital capabilities, is designed to address the increasingly distributed requirements of enterprise, cloud and AI customers.
“Today’s AI infrastructure build-out is driving demand for our connected environments in strategic locations around the world,” he said.
