GigaCapacity
Free tool
Should a buyer build or lease AI data center capacity?
Updated June 14, 2026

Build vs lease

Build vs Lease AI Capacity Calculator

Use this calculator to compare an owned build against a leased AI capacity path across NPV, time-to-power value, utilization risk, and term flexibility.

The answer is fit-based: building can win for controlled long-term load, while leasing can win when speed, phased growth, or demand uncertainty matters more than asset control.

How to use this calculator

Use before issuing a build RFP, signing a lease LOI, or deciding whether near-term GPU demand justifies permanent owned capacity.

  1. 01

    Enter the buyer scenario

    Start with the editable build vs lease inputs and replace defaults with current quotes, tariffs, engineering values, or buyer assumptions.

  2. 02

    Review the modeled outcome

    Use the modeled outcome, decision checks, and copied brief to compare the scenario against buyer constraints.

  3. 03

    Verify the result before acting

    Check the caveats, source table, and related guides before treating the output as a quote, bid, or final site decision.

Calculator inputs

Scenario defaults are editable. Replace them with current quotes, utility tariffs, tax counsel inputs, or engineering values before relying on the output.

Lease-favored

Modeled outcome

Lease is favored on this scenario because speed value and flexibility outweigh the owned-build NPV advantage.

Build capex

$224M

Build NPV

$245M

Lease NPV

$218M

Speed value

$518M

Lease advantage

$545M

Decision checks

  • Build NPV: $244.65M
  • Lease NPV: $218.26M
  • Speed value: $518.4M

How does the build vs lease model work?

The model calculates an owned-build NPV from upfront capex, owner opex, utilization, and residual value, then compares it with leased capacity payments over the same term. It separately values the months saved by leasing when leased capacity can be energized before an owned build. The recommendation weights NPV difference, speed value, utilization risk, and term length so the buyer can see which operating model fits the scenario.

What should buyers verify before using the result?

  • Lease rates can include different boundaries for power, cooling, remote hands, cross-connects, and utility pass-throughs.
  • Owned-build economics depend heavily on residual value and whether the buyer can use the asset beyond the initial workload.
  • Time-to-power value should be set to zero if the workload has no near-term revenue or strategic deadline.

Related guides and tools

Which sources support this build vs lease model?

SourceUse in this toolLink
TierPoint data center cost calculatorCompetitive calculator pattern for build-versus-colocation decisions.Open
Expedient data center build vs buy calculatorCompetitive calculator pattern for build-versus-buy modeling.Open
Digital Realty build vs buy reportStrategic factors in build-versus-buy decisions.Open
CoreSite data center costs build or lease guideBuild-versus-colocation factors for enterprise buyers.Open
CBRE North America Data Center Trends H2 2025Market supply, vacancy, and leasing pressure across North America.Open
JLL 2026 Market Outlook for Global Data CentersGlobal market outlook, cost, site selection, and power context.Open
Uptime Institute Global Data Center Survey 2025Data center efficiency and operations benchmark context.Open
U.S. EIA Electric Power MonthlyElectricity price and sales context for power cost assumptions.Open

FAQ: Should a buyer build or lease AI data center capacity?

Building tends to fit when the buyer has stable long-term demand, internal operating capability, site control, power certainty, and a reason to own the asset after the first term.

Leasing tends to fit when speed, phased demand, geographic optionality, or execution risk matters more than owning the facility.

Normalize critical kW, included cooling type, metered power treatment, escalators, cross-connect charges, expansion rights, SLA credits, and make-ready costs.