Digital Realty adds programmable AI interconnection controls

Digital Realty has introduced programmable interfaces for its interconnection platform, targeting enterprises deploying private AI infrastructure across its network of over 800 data centers. The ServiceFabric Model Context Protocol (MCP) provides standardized controls for connectivity, telemetry, identity management, and operational policies, allowing organizations to run AI workloads near data sources while maintaining governance and flexibility in partner selection and geographic placement.

The move reflects broader industry efforts to address the infrastructure challenges of AI adoption. While Digital Realty does not position the launch as a complete solution for private AI, it aims to simplify one critical layer: the physical and network infrastructure connecting enterprise data, AI systems, cloud providers, and colocation facilities. The protocol is designed to scale across diverse environments, from traditional enterprise workloads to high-density AI clusters.

What changed

ServiceFabric MCP exposes programmable interfaces for four core functions: network connectivity, real-time telemetry, identity verification, and operational policies. Enterprises can now automate interconnection workflows rather than relying on manual provisioning or static configurations. The protocol is available across Digital Realty’s global footprint, which spans major AI hubs in North America, Europe, and Asia-Pacific.

The launch follows a series of recent expansions by Digital Realty, including a 10 MW computing cluster for Mistral AI at its Paris South campus and the deployment of ORCA Computing’s quantum hardware in its London Innovation Lab. These developments suggest a strategic focus on supporting emerging compute workloads alongside traditional colocation services.

Industry context

The timing of the announcement coincides with heightened scrutiny of data center power consumption and infrastructure demands. Digital Realty’s 2025 Impact Report, released earlier this month, claims 93% of its global power consumption is now covered by renewable energy, a figure that arrives as regulators and utilities increasingly question the environmental and grid impacts of AI-driven data center growth.

Competitors are also addressing AI infrastructure challenges through partnerships. Schneider Electric and Foxconn, for example, announced a collaboration this week to integrate power, cooling, and rack systems for faster deployment of high-density AI data centers. Unlike vendor-specific alliances, Digital Realty’s approach emphasizes interoperability and policy control, positioning its platform as a neutral layer for enterprises managing multi-cloud and multi-partner AI environments.

What to watch

Adoption of ServiceFabric MCP will likely depend on how effectively Digital Realty integrates the protocol with existing AI infrastructure providers. Early use cases may emerge in regulated industries, such as finance and healthcare, where data sovereignty and policy control are critical. The company’s ability to demonstrate measurable improvements in deployment speed and operational consistency will also influence enterprise uptake.

Separately, the broader data center industry faces growing pressure to reconcile AI-driven demand with sustainability commitments. Digital Realty’s renewable energy claims will be tested as AI workloads increase power consumption across its facilities. Observers will watch whether similar programmable interconnection tools become an industry standard or remain a proprietary offering.