If you're coming from the First Principles Brief, you already know what Digital Core does and why the business exists. This is where the analysis deepens.
The Brief ended on one distinction and five questions built around it.
Portfolio growth and per-unit value are not the same thing, and the Brief argued that Digital Core could keep enlarging its asset base without making each unit meaningfully more valuable.
What follows does not re-describe the business. It takes that distinction and tests it against evidence the Brief did not use. That evidence is the FY2023 to FY2025 annual reports, the 1H2026 interim results to 30 June 2026, and a large related-party transaction the sponsor announced on 12 August 2026.
The first half of the story confirms the concern.
Three and a half years in, the distribution per unit has still not grown.
The second half complicates it. Management now points to a re-leased anchor asset and an accretive recycling of the portfolio, and guides to double-digit distribution growth in 2027.
The decision comes down to whether to believe that turn before it arrives, and at what price.
Check out the First Principles Brief here:
Ryan Gallinera
Digital Core owns 11 freehold data centres across the United States, Canada, Germany and Japan, valued at about US$1.9 billion, with roughly 1.02 million square feet of in-service net rentable area, in-service occupancy of 97.3% and a weighted average lease expiry of about 4.3 years as at 30 June 2026.1 That is the asset. The competitive question is narrower than the asset base. It is how much of the demand for digital infrastructure the landlord can hold onto, rather than the chipmaker, the cloud platform or the tenant that operates the servers.
The answer starts below the floor. A data centre is not valuable because it is a large building. It is valuable because of what feeds it.
The binding constraint in this industry is electricity, not concrete. In the major markets Digital Core operates in, securing new grid capacity has become slow and, in several metros, effectively rationed. Vacancy in the top North American markets fell to a record low in early 2026, and Northern Virginia, the sector's largest hub and one of Digital Core's core markets, ran at about 0.3% vacancy.2 A facility that already holds committed, energised power in one of those markets owns something a competitor cannot simply build alongside it.
This is what separates a data-centre asset from an ordinary industrial property. Warehouse supply responds to demand within a building cycle. Powered data-centre capacity does not, because the bottleneck sits with the grid rather than with the developer. Digital Core's portfolio is fully leased or nearly so, at 97.3% in-service occupancy, in exactly the markets where that scarcity is most acute. The rent is protected less by the walls than by the near-impossibility of a tenant finding equivalent powered space nearby at short notice.
A working data centre requires several scarce inputs assembled at once. Suitable land, planning consent, grid power, cooling, network fibre and physical security all have to come together on the same site. Each one lengthens the development timeline, and any one of them can stall a project. Industry analysts describe an infrastructure investment supercycle, with global capacity forecast to nearly double from about 103 GW to 200 GW by 2030 and roughly three-quarters of the construction pipeline pre-committed to tenants before it completes. Demand is running ahead of what supply can deliver.
For an owner of existing, stabilised, powered capacity, that slow supply response is the moat. The scarce thing is not the ambition to build. It is permitted, powered, connected capacity that already exists.
Once a tenant installs its own equipment inside a facility, leaving is expensive and operationally risky. Migrating live infrastructure means downtime, re-cabling, re-testing and a window of elevated failure risk, all for a service that customers expect to run without interruption. That friction is why data-centre leases run for years and why tenants tend to renew rather than move.
The friction shows in the rent roll. Close to 80% of Digital Core's rental income comes from investment-grade tenants, dominated by large cloud and technology companies that sign multi-year leases and rarely default.3 In the first half of 2026 the manager signed about US$5 million of new and renewal leases at a positive 25% cash rental reversion, and less than 10% of the portfolio now expires in any single year through 2029.4 For the landlord, that is a high-quality, durable income stream, and the near-term expiry risk has been reduced. It is also the source of the moat's most important limit, which the next section reaches.
The Linton Hall property in Northern Virginia lost its tenant when a lease expired, and the income stopped. That is the exact risk the switching-cost story is supposed to prevent, and it happened. What followed is the more useful evidence. In January 2026 the manager signed a ten-year agreement with an investment-grade global cloud service provider to take the entire facility, commencing on 1 December 2026 at about a 35% uplift on the previous net rent. The lease is expected to generate roughly US$14.8 million of annualised net property income at the asset level, or about US$13.3 million at Digital Core's then 90% share, and on commencement portfolio occupancy improves to about 98%, the investment-grade share of rent rises to about 83%, and the portfolio lease expiry extends to about 5.1 years. The vacancy was real and the asset is still in redevelopment, but the market rent had risen enough that the replacement lease comes in materially higher.
That episode is the moat working and failing at the same time. It failed to retain the original tenant. It succeeded in re-pricing the space upward, because powered capacity in that market is scarce. A landlord will take that trade. It confirms both the pricing power and the asset-specific risk that no amount of sector demand removes. It also matters for what comes later, because Linton Hall's return to full rent in 2027 is the single largest driver of the growth management now promises.
First Principles Analyses and Research Snapshots are included with Full Research Access.
