For most of construction history, the critical path of a building ran through the building: foundations, structure, envelope, systems, finishes. On campus-scale data center programs it increasingly runs elsewhere, through interconnection studies, substation outages, transformer production slots and switchgear test bays far from the site fence.
Schedules have been slow to follow. Many still show utility power as a single milestone and long-lead electrical equipment as a line in a procurement log. Both choices put the program's most consequential risks outside the logic that is supposed to forecast them.
Power is the new foundation
Utility capacity for a large campus is rarely waiting to be connected. It has to be studied, agreed, designed, procured, built and energized, often through a sequence that is longer than the building program itself. Treating that sequence as a date the owner hopes for, rather than a chain of activities with owners and dependencies, is the single most common structural weakness we see on these programs.
A credible integrated master schedule models the utility path with the same rigor as the shell and core: study phases and their review windows, agreement execution, high-voltage design, equipment procurement, construction, outage windows and energization, each tied to the work it enables.
The factory floor is on the critical path
Transformers, medium-voltage switchgear, generators and UPS systems can carry lead times that dominate the program. The controlling milestones are no longer only on site. They are design freeze for each equipment package, production slot confirmation, factory acceptance testing, shipping windows and the site's readiness to receive and set heavy equipment.
- Design freeze dates that release each equipment order
- Factory slot confirmations and the penalties for missing them
- Factory acceptance tests and owner witness requirements
- Heavy-haul logistics, permits and setting windows
- Energization sequences that unlock commissioning
When these milestones live inside the schedule, a late drawing approval shows up as a forecast movement within one update. When they live in a spreadsheet, the same late approval tends to surface when the equipment does not arrive.
In a data center, a missed date is not an inconvenience. It is contracted capacity a customer cannot use.
Commissioning is a phase, not a finish line
Level 1 to 5 commissioning, ending in integrated systems testing, can take a material share of the final months. When it is planned late and by area, construction finishes in the wrong order for the systems that need to be tested, and the last phase compresses into overtime and rework. The stronger programs define commissionable systems during design and sequence construction completion to suit them.
P80 for capacity, not P50
Because capacity dates are often sold before the building is finished, the forecast that matters is not the most likely date but the date the program can defend. We encourage owners to commit externally to a P80 date from a risk-loaded schedule, manage internally to the P50, and treat the gap between them as contingency to be protected, not consumed.
Repeatability as an advantage
Every building on a campus is a learning cycle. Standard logic templates, consistent activity coding and captured productivity let each phase start from the last one's lessons rather than from a blank page. Over several buildings, that turns schedule performance into a compounding asset.
What owners should ask for
Owners of large campuses should expect controls that run from grid to rack: an integrated schedule that connects utility, procurement, construction and commissioning; long-lead equipment visible in the logic; and reporting that shows where the risk actually sits, even when it sits a thousand miles from site.
