AI Data Center & HPC Lease Valuation Calculator
Model an HPC/AI data center colocation lease the way a real estate desk would — critical-IT megawatts, lease rate, term, and NOI margin into an exit-cap discounted cash flow. Get equity value, implied value per share, and cap-rate sensitivity. Prefill the structural terms from SEC-sourced deals for the US bitcoin miners and neoclouds building this capacity.
HPC Lease CalculatorLandlord
Capacity
Gross MWassumption
PUEassumption
Critical IT MWderivedGross MW ÷ PUE
Contract
Lease Rateearns
$/kW/mo
per critical IT kWLease Termassumption
years
Rent Escalatorassumption
%/yr
blank = flat · typical 2–3.5NOI Marginassumption
%
typical 80–90 (tenant pays power)Capital & exit
Datacenter Capexassumption
$M/crit MW
typical $9–15MExit Cap Rateassumption
%
IG hyperscale 5.25–6.5 · colo 6–7.5 · implies 15.4xWACCassumption
%
discount rateProbabilityassumption
%
Shares Outstandingassumption
M
Probability-Adjusted Equity Value$1.26B
Implied Value / Share$4.89
Critical IT MW214
Year-1 Rent$372.9M
Year-1 NOI$316.9M
Exit-Year NOI Y10, escalated$316.9M
Total Capex-$2.57B
Yield on Cost12.3%
Discounted
PV of Contracted NOI 10 yrs @ 10.0%$1.95B
PV of Exit Value NOI ÷ cap, discounted$1.88B
Less: Capex-$2.57B
Equity Value$1.26B
Undiscounted (competitor-comparable)
Stabilized Value Y1 NOI ÷ cap$4.88B
Value Created vs. build cost$2.30B
Exit cap sensitivity — value / share
| Exit cap | 4.5% | 5.5% | 6.5% | 7.5% | 8.5% |
|---|---|---|---|---|---|
| Value / share | $8.15 | $6.23 | $4.89 | $3.92 | $3.17 |
Discounting costs $1.05B against the undiscounted stabilized view — that gap is what a cap-rate-only model hides.
Questions
How do you value a data center colocation lease?
Start from critical-IT capacity (gross MW ÷ PUE) and the lease rate in $/kW/month to get contracted rent, then apply the NOI margin — in a triple-net colocation lease the tenant pays power, so the margin is high (typically 80–90%). That produces annual net operating income. The lease value is the present value of that NOI stream over the contract term plus the discounted value of the asset at exit (exit-year NOI capitalized at an exit cap rate), less the capex to build. Dividing by shares outstanding gives an implied value per share.
What is an exit cap rate?
A capitalization rate converts a single year of stabilized NOI into an asset value: value = NOI ÷ cap rate. The exit cap rate is the rate you assume a buyer would pay at the end of the hold — so the terminal value in the DCF is the exit-year NOI divided by that cap rate. A lower cap rate implies a higher price (a 6.5% cap is a ~15.4× multiple on NOI); investment-grade hyperscale leases trade tighter (5.25–6.5%) than merchant colocation (6–7.5%).
Why discount the cash flows instead of just capitalizing NOI?
Capitalizing NOI (NOI ÷ cap rate) treats the asset as if it were already stabilized and owned free and clear today. It ignores the time value of a 15–20 year contract, the capex you spend up front, and the fact that the exit value arrives years from now. A DCF discounts each year of NOI and the terminal (exit) value back at your cost of capital (WACC), then nets out capex — so it captures the drag from build cost and the delay to the exit. The gap between the two is real money a cap-rate-only model hides.
What is critical IT MW and how does PUE affect it?
Gross MW is the total power drawn by the facility; critical-IT MW is the portion delivered to the servers themselves. Power usage effectiveness (PUE) is the ratio of total to IT power, so critical-IT MW = gross MW ÷ PUE. Lease rates are quoted per critical-IT kW, so PUE directly scales the revenue base: a 250 MW site at PUE 1.4 leases ~179 critical-IT MW.
Where do the prefilled deal figures come from?
Loading a tracked deal — its capacity and term straight from SEC filings, plus our modeled lease rate and NOI margin — is a subscriber feature: a Dimetrics subscription (from $29/mo) auto-fills every field for the tracked US bitcoin miners and neoclouds. On the free tier the calculator starts at sensible defaults and you enter your own assumptions for any company or prospective deal — the math is identical, you just source the inputs yourself.