Digital Realty has introduced ServiceFabric MCP, a new AI-native control layer that extends its global interconnection platform with programmable infrastructure capabilities for enterprise AI deployments. Available immediately across more than 800 Digital Realty and partner-connected data centers, the platform adopts the emerging Model Context Protocol (MCP) standard to provide AI agents and applications with standardized interfaces for discovering, provisioning, monitoring, and managing infrastructure resources. The company positions the launch as a key component of its broader Foundation for AI strategy, which focuses on enabling large-scale enterprise AI deployments across distributed environments.
The new platform builds on Digital Realty’s AI Private Exchange (AIPx) architecture, which incorporates policy and orchestration technologies designed to support private AI environments. ServiceFabric MCP introduces programmable controls across four operational domains: intent-based connectivity design and provisioning, real-time topology and telemetry discovery, identity and security management, and operational integration with enterprise software platforms. According to Digital Realty, the platform enables AI systems to securely interact with networking, colocation, and interconnection resources while maintaining private Layer 2 and Layer 3 connectivity with authentication and access controls.
Digital Realty said the technology is currently operating within its own infrastructure environments and is being validated across customer deployments and ecosystem partnerships. The company highlighted collaborations with ePlus, Lenovo, and Dell, alongside infrastructure powered by NVIDIA and AMD technologies. Healthcare AI startup See All AI cited its use of Digital Realty’s ServiceFabric and the company’s Borton campus infrastructure to support NVIDIA DGX B200 systems for medical imaging workloads. Looking ahead, Digital Realty said ServiceFabric MCP represents the first programmable layer of a broader architecture that could eventually encompass power management, space allocation, inventory systems, partner ecosystems, and sovereign AI deployment requirements.
• ServiceFabric MCP adopts the emerging Model Context Protocol (MCP) standard for AI infrastructure control.
• Available across more than 800 Digital Realty and third-party connected data centers.
• Built on Digital Realty’s AI Private Exchange (AIPx) architecture.
• Supports intent-based provisioning, telemetry discovery, security controls, and operational integrations.
• Integrates with platforms including Slack, Microsoft Teams, Splunk, and Datadog.
• Supports multi-cloud, colocation, bare-metal, and hybrid AI deployments.
• Designed to support private AI environments without locking customers to a single AI model or cloud provider.
“ServiceFabric MCP extends the foundation of AIPx with programmable controls and agent-ready interfaces, and our patent position reflects the long-term investment we’ve made in this architecture,” said Chris Sharp, Chief Technology Officer of Digital Realty.
🌐 Analysis
Digital Realty’s announcement reflects a broader industry effort to make physical AI infrastructure programmable and accessible to AI agents through standardized interfaces. MCP, originally gaining traction within the AI software ecosystem, is increasingly being evaluated as a mechanism for connecting AI systems directly to infrastructure resources, allowing automated provisioning, monitoring, and orchestration. Similar trends are emerging across cloud providers, networking vendors, and infrastructure software companies seeking to expose infrastructure functions through AI-friendly APIs.
For Digital Realty, ServiceFabric MCP also strengthens its competitive positioning against hyperscale cloud providers and AI-focused infrastructure operators. Rather than competing on AI models or GPU services, Digital Realty is focusing on the physical layer requirements of enterprise AI—including power density, cooling, interconnection, data sovereignty, and multi-site deployment flexibility. The launch complements recent Digital Realty investments in expanding hyperscale capacity, increasing ownership of Teraco in Africa, and developing AI-ready infrastructure platforms that span colocation, connectivity, and interconnection services globally.
How MCP Could Transform Network Operations Why the Model Context Protocol matters for AI-driven networking, automation, and infrastructure operations • Updated June 2026 | |
| Executive View: MCP is not a replacement for networking protocols such as BGP, NETCONF, RESTCONF, SNMP, or gNMI. Instead, MCP acts as an AI-facing integration layer that allows AI agents and applications to securely discover infrastructure, retrieve telemetry, invoke tools, automate workflows, and interact with operational systems through a common interface. | |
| Intent-Based Operations | An engineer or AI agent could request: “Connect this AI training cluster to storage with low latency and redundant paths”. MCP can translate that request into orchestrated actions across inventory, connectivity, provisioning, and policy systems. |
| Infrastructure Discovery | Provides AI systems with structured visibility into network topology, available capacity, ports, cross-connects, cloud on-ramps, colocation assets, and data center resources. |
| Real-Time Telemetry | Allows AI agents to access performance data such as latency, throughput, packet loss, congestion, link utilization, optical health metrics, power consumption, and environmental data. |
| Service Provisioning | Enables AI-assisted deployment of VLANs, VRFs, EVPN services, cloud interconnects, private AI exchanges, bandwidth upgrades, and network segmentation policies. |
| Fault Isolation | AI agents can correlate alarms, topology data, maintenance records, telemetry streams, and recent configuration changes to identify probable root causes and accelerate mean-time-to-resolution. |
| Policy & Security Controls | Supports role-based access, identity management, geographic restrictions, sovereign deployment requirements, compliance controls, and change-management policies. |
| NOC & IT Operations Integration | Connects AI workflows with operational tools including Splunk, Datadog, ServiceNow, Microsoft Teams, Slack, observability platforms, ticketing systems, CMDBs, and enterprise knowledge bases. |
| Human-in-the-Loop Automation | Production-grade deployments can require approvals before executing changes, preserving auditability, rollback capabilities, compliance requirements, and operational governance. |
| AI Infrastructure Operations | Particularly relevant for AI factories and GPU clusters where networking, storage, power, cooling, and workload placement increasingly require coordinated orchestration across multiple domains. |
| Industry Significance | The long-term opportunity: MCP may become a common interface layer that allows AI systems to interact with networking, data center, cloud, and operational infrastructure without requiring custom integrations for every platform. For operators, this could create a new generation of AI-assisted NOCs where infrastructure becomes directly accessible to intelligent software agents while retaining human oversight. |