To rank a data center site shortlist, first remove any site that fails a gating test, then score the survivors on weighted criteria, with deliverable power and time to energization weighted highest. A site that scores well on everything except power should not outrank a site with a credible power date. Ranking this way gives the buyer a defensible order and a clear reason for each site's position.
Step 1: Apply gating tests
A gating test is pass or fail. Any site that fails is removed from the ranking, no matter how attractive it looks elsewhere.
Common gates for data center sites:
Power path exists: a utility service request, load study, or credible on-site or behind-the-meter plan, not just proximity to a line
Site control is achievable: an owner willing to sign an option, LOI, or purchase agreement on workable terms
Use is permissible or permittable: zoning allows data centers, or a realistic rezoning or special approval path exists
Buildable area fits the campus: after flood zones, wetlands, buffers, and setbacks
No known fatal flaw: for example a local moratorium, a title defect, or contamination with no remediation path
Step 2: Score the survivors
Score each remaining site from 1 to 5 on each criterion, multiply by the weight, and add up the total.
The weights below are example weights for a site meant to take load in phases. Adjust them to the buyer's priorities.
| Criterion | What a 5 looks like | Example weight |
|---|---|---|
| Deliverable power and date | Load study done, MW and energization window documented | 30% |
| Time to energization | Shortest credible path to first load | 15% |
| Entitlement path | By-right use or approval in hand | 15% |
| Site control terms | Signed or ready-to-sign option with fair terms | 10% |
| Fiber | Multiple carriers or routes nearby, clear lateral path | 10% |
| Water and cooling fit | Supply confirmed for the planned cooling design | 10% |
| Total cost to ready | Land, upgrades, and site work within budget | 10% |
Step 3: Apply tie-break and override rules
Power beats everything else at a tie. If two sites score within a few points, rank the one with the more certain power date first.
Unverified scores are capped. If a criterion is scored from desktop data only, cap it at 3 until it is confirmed.
Expansion counts. For phased campuses, prefer the site with a credible path to the next tranche of MW.
Community risk is real. A site with visible local opposition or a pending moratorium drops one rank until resolved.
What each rank should come with
A ranking is only useful if the buyer can see why. For each site, include:
The total score and the score on each criterion
The evidence behind each score, with source and date
What is confirmed and what is still assumed
The next three actions to move the site forward
The single biggest risk and how to retire it
Common mistakes when ranking data center sites
Ranking on acreage and price first. Cheap land with no power date is not a good site.
Treating proximity to a substation as capacity. Distance is not deliverability.
Mixing verified and assumed data without labels. It makes weak sites look strong.
Never re-ranking. Utility studies, owner terms, and local politics change. Re-rank when new evidence lands.
Ranking too many sites. A shortlist works best when it is short enough to act on.
How neoclouds and developer sites teams use a ranked shortlist
A ranked shortlist turns a site search into a decision. Sites teams use it to decide where to spend on options, studies, and entitlement work. Neoclouds use it to compare speed to power across markets. In both cases, the rank should change what the team does next week.
How Build helps
Build sources data center sites, validates them against power, entitlement, fiber, water, and site control, and bundles the services needed to take the top site from shortlist to usable land. The client gets a ranked list with the evidence behind every score, and when a deal stalls, another validated site to move on.