A data center lease exhibit that defines capacity three different ways deserves careful reading before the expansion model is trusted

Diligence on a stabilized fund's largest data center asset can turn on a single exhibit. The marketing materials might describe the asset as a 32 MW facility. The lease's Exhibit C could define "Contracted Capacity" as 24 MW of critical IT load. Exhibit D, the utility schedule, might reference a 36 MVA service. The engineering report might use a "design capacity" of 32 MW, footnoted as assuming N+1 rather than 2N on the UPS plant. So the number that generates rent is 24, the number the wires can carry is roughly 29 at a reasonable power factor, and the number in the deck is 32. None of that is necessarily dishonest on its own, but it changes how an expansion story should be read. If a fund's model shows lease-up of a remaining "8 MW" at market rates starting in year three, and that 8 MW only exists when measured from design capacity, the real expansion block against contracted capacity and actual service can be closer to 4 MW, and even that may need additional cooling capacity that the engineering report has only placeholder pricing for. Curtailment language deserves the same scrutiny. Most such provisions carve out utility events on the theory that generators exist for exactly that purpose. When a lease's abatement threshold counts utility events without a carve-out, the model's assumption of zero abatement exposure is worth challenging directly. The right move in a file like this is to pull the interconnection agreement and several years of outage logs and underwrite both the expansion block and the curtailment exposure on the more conservative reading until the documents say otherwise.