On September 15, Digital Realty (NYSE:DLR) made ServiceFabric MCP available, a software layer that lets AI agents design, monitor and troubleshoot network connections across more than 800 data centers, including third-party sites. That nudges a real estate company toward becoming a control panel for enterprise AI. It arrives after second quarter results reported on July 23, when Core FFO per share, the company’s preferred earnings yardstick, excluding net promote rose to $2.13 from $1.87 a year earlier. Here is what the launch does, and what it has yet to prove.
Plumbing With a Brain
The pitch is that enterprise AI needs more than servers. It needs power, cooling, and sovereign placement that software can control. ServiceFabric MCP handles four jobs: designing and provisioning connections, spotting capacity and watching live network health, managing access through OAuth 2, and handing troubleshooting to agents with links into chat and monitoring tools. It is also open by design. Customers do not have to live only in Digital Realty buildings or commit to a single AI model. An IDC research VP argues that public cloud interfaces alone cannot give enterprises enough control over data movement and policy, which favors providers that pair global reach with programmable interconnection. Digital Realty runs the system on its own AI workloads, and See All AI, a medical imaging developer, leans on the Boston campus and ServiceFabric to move large datasets quickly and securely.
The financial engine underneath is running hot. Renewal leases in the second quarter were signed at rates 25.4% higher on a cash basis, which shows customers will pay more to stay put. Signed leases waiting to start added up to a $1.9 billion backlog of annualized base rent at 100% share, so future revenue is already lined up. Management responded by lifting its 2026 Core FFO per share outlook, excluding net promote, to $8.15 to $8.20.
Big Checks, Bigger Questions
Start with what the launch has not shown yet. Digital Realty itself calls MCP an emerging standard, and ServiceFabric MCP is still being validated across internal, enterprise, and partner deployments. The announcement puts no dollar figure on what it could add to revenue, and the company describes it only as the first programmable surface of a larger architecture that may later stretch into space, power and inventory. Until customers pay for this layer, it is a promising idea more than a line item.
Then there is the bill for the physical side. Digital Realty carried about $18.6 billion of debt at June 30, 2026, and its 2026 development spending outlook, net of partner contributions, now sits at $4.25 billion to $4.75 billion. The outlook also assumes new long-term debt priced at 4.5% to 5.5%, up from the earlier 4.0% to 4.5%. To help pay for growth, the company has sold roughly 13.5 million shares this year for about $2.5 billion, which spreads future earnings across more owners. And the quarter’s headline flattered a bit: Core FFO per share of $2.65 included a $188 million net promote, so the cleaner figure is the $2.13 mentioned above.
Crowd Grows, Price Climbs
Hedge funds holding Digital Realty climbed to 73 from 46 in the prior quarter, so professional money is arriving rather than leaving. The forward P/E stands at 67.11 as of September 18. That means investors are paying a steep price today for earnings still to come. A multiple like that assumes growth keeps delivering, and it leaves little cushion if a quarter disappoints. Funds are buying anyway, which sets the crowd’s confidence directly against the price.
Promise Versus Proof
ServiceFabric MCP shows where Digital Realty thinks enterprise AI spending is heading, but a launch is not the same as adoption. The tension is that the physical business is growing fast today while the software layer has yet to prove it can earn its own keep. Bulls need customers actually using the agent-ready controls, and those controls pulling more workloads onto the platform. The bear worry is simpler: can the company fund its building spree without leaning ever harder on new shares and pricier debt? Over a 3- to 5 year stretch, usage will settle that debate, not announcements.
READ NEXT: 10 Best Future Stocks to Buy Under $10 and 12 Best Performing Semiconductor Stocks to Invest In.