Strategic Asset or Dead Weight: A Rubric for Deciding What Your Data Center Is Really Worth

Every owned data center in the American enterprise portfolio arrived for a defensible reason — a compliance mandate, a latency requirement, a lease signed before cloud economics matured. The problem is not that these facilities exist. The problem is that few organizations periodically ask, with real rigor, whether the facility still earns its keep.

A data center is either a strategic asset or ballast. Ballast is not useless — it stabilizes the ship. But ballast also adds weight, consumes fuel, and limits how fast you can turn. The distinction matters most now, as hybrid architectures force enterprises to make placement decisions across owned, colocation, and public cloud capacity simultaneously.

Why the Question Gets Deferred

Owned facilities rarely face the scrutiny applied to cloud spend. There is no monthly invoice line item that spikes and triggers a review. Costs are amortized, embedded in facilities budgets, and spread across depreciation schedules long enough to make any single year look tolerable. Meanwhile, the teams operating the facility have every professional incentive to keep it running well — which is not the same as keeping it running economically.

The result is a quiet asymmetry: cloud commitments get quarterly reviews, while a 40,000-square-foot building with aging chillers and 30 percent rack utilization goes unexamined for a decade.

A Five-Factor Scoring Rubric

We recommend a disciplined annual scoring exercise, rated one to five on each dimension, with anything below 18 total points triggering a formal strategy review.

1. Power economics (weight: highest). Compare your all-in cost per kilowatt-hour — including demand charges, PUE overhead, and cooling — against regional colocation and hyperscaler effective rates. Facilities in markets with rising industrial power costs, or on grids exposed to volatile demand pricing, lose ground every year. A facility paying more than a 25 percent premium over comparable colocation is carrying structural ballast.

2. Productive utilization. Measure not floor space but usable, properly conditioned rack capacity actually supporting production workloads. Below 40 percent, fixed costs are being amortized over an asset base that doesn't justify them. Above 75 percent without a modernization plan, you face a capacity cliff instead.

3. Technical currency. Score the age and efficiency of power distribution, cooling, and network fabric against what a new build or modern colocation provider would offer. A facility that cannot support high-density compute — including AI-adjacent workloads — is not merely outdated; it is excluded from future architecture decisions.

4. Workload gravity. How many production applications are genuinely pinned to the facility by latency, data residency, or licensing constraints? Be honest. Many "pinned" workloads are pinned by habit and migration cost, not requirement.

5. Exit economics. The most overlooked factor. What would it actually cost to leave — decommissioning, lease buyouts, data migration, contract wind-downs? High exit costs do not justify staying, but they must be priced into the timing of any decision. Deferring the choice makes the eventual exit more expensive, not less.

Interpreting the Score

Scores above 30 generally indicate a facility worth modernizing — invest in density, efficiency upgrades, and possibly subleasing excess capacity. Mid-range scores suggest a hybrid path: migrate the movable workloads, retain a hardened core, and shrink the footprint deliberately. Scores below 18 mean you are operating ballast, and every quarter of delay converts sunk cost into stranded cost.

The rubric's real value is not the number. It is that it forces a conversation between facilities, infrastructure, and finance teams who otherwise each hold one-third of the truth. An enterprise that cannot assemble those three perspectives in a room has a governance problem before it has a facilities problem.

Decide on a Schedule, Not on a Crisis

The worst time to evaluate a data center is during a failure, a lease expiration, or an acquisition. Put the assessment on the same calendar as your cloud commitment reviews. Owned capacity deserves the same discipline as contracted capacity — no more, no less.

The anchor metaphor is apt here: an anchor is valuable when it holds you in place and costly when it prevents you from moving. Knowing which one your facility is, on an annual basis, is a small exercise with a very large balance sheet implication.