AXG and Digital Realty Partner to Enable GSIs & MSPs to Deliver Hosted Private Infrastructure as a Service in More Than 30 Countries

AXG and Digital Realty formed a partnership to provide lifecycle-managed private AI, private-cloud and edge infrastructure as a service across 31 markets in North America, EMEA, APAC and Latin America. The offering combines Digital Realty's data-center, interconnection and sovereign-ready colocation platform with AXG's infrastructure financing, deployment and lifecycle management, using one recurring monthly charge with no upfront customer infrastructure capex. The partnership targets enterprise and government demand for private AI, edge computing, cloud repatriation and neo-cloud/GPU-as-a-service capacity.

Analysis

The economic value to DLR is not the partnership announcement itself but whether AXG converts channel access into contracted power commitments. The model lowers procurement friction for GSIs/MSPs and can improve DLR’s mix toward higher-value, interconnected deployments; however, DLR remains primarily exposed to lease commencement timing, booked-MW conversion, and the spread between escalators and power/capex costs. Treat this as a modest pipeline-enablement signal rather than an earnings-revision event until management discloses incremental signed capacity, lease terms, or materially improved utilization.

The more important second-order effect is competitive: private AI and repatriation favor wholesale/colocation operators with global footprints and sovereign-data options, but they also reduce the addressable workload pool for hyperscalers over time. DLR’s direct peers EQIX, COR and, to a lesser extent, AMT benefit from the same enterprise demand trend; DLR’s differentiation depends on winning large multi-region requirements where its global capacity and interconnection can outweigh EQIX’s denser enterprise ecosystem. GPU supply and enterprise AI-utilization uncertainty remain the bottleneck: dedicated capacity is attractive only when customers can sustain utilization sufficiently to justify long-duration commitments.

Over the next 1-3 months, watch DLR’s bookings commentary, pre-leasing of AI-ready capacity, and any evidence that channel partners are converting to multi-market contracts. Over 6-18 months, a sustained cloud-repatriation cycle could support occupancy, pricing and development yields, but aggressive new builds by DLR, EQIX, private infrastructure funds and neo-cloud providers could cap rent growth. The thesis is falsified if DLR reports weak signed-MW bookings, rising concessions, delayed customer power turn-ups, or development yields deteriorating as power procurement costs rise.

Key Decisions for Investors

• Maintain a modest long DLR versus short EQIX only if DLR continues to trade at a material FFO multiple discount and upcoming results show accelerating signed MW or backlog conversion; target a 3-6 month normalization trade, with exit if DLR booking momentum fails to improve for two reporting periods.
• Do not add directional DLR exposure solely on this release. Set an event-driven alert for disclosed AXG-linked contracted capacity, initial lease duration and power density; without those data, revenue and FFO sensitivity cannot be credibly underwritten.
• For AI-infrastructure exposure, favor a basket approach—long DLR and EQIX against a small short in a broad office REIT proxy such as BXP—over a concentrated DLR bet. The relative trade isolates resilient digital-infrastructure demand from broader commercial-real-estate refinancing risk over 6-12 months.
• Monitor DLR’s development yield, interest expense trajectory and power-cost pass-through at the next earnings release. A decline in stabilized yield or a material rise in incentives would indicate that AI capacity competition is transferring economics to customers rather than shareholders.