BitcoinMiningStock.io is now Ziven.io. Here's why.

Colocation Deal Simulator & Pricing Calculator

How a colocation deal creates value

A colocation deal works like this: the operator is the landlord. It provides the power and cooling infrastructure, and the tenant brings the servers. The operator builds at a high return (yield on cost), but the market values the finished asset at a lower return (cap rate). That gap is the value created.

Step 1

The deal

What the company announced. Type the headline terms.
Deal name optional
Critical-IT MW
megawatts
Contract value
total over the term · "19B" or "19000000000"
Term
years
Escalator
%/yr · blank = flat ·
Year-1 annual rent
$ per year · e.g. 725M
Step 2

Your assumptions

Nobody knows these for sure; they're yours to flex. The exit cap rate matters most, so test a range with the slider in the results. Enter the deal above to set your assumptions.
Bull
Exit cap rate
%
NOI margin
%
Build cost / MW
$ per MW
Base drives the results
Exit cap rate
% ·
NOI margin
% · (tenant pays power)
Build cost / MW
$ per MW ·
Shares outstanding optional
shares · enables per-share view
Bear
Exit cap rate
%
NOI margin
%
Build cost / MW
$ per MW

Bull and Bear derive from Base (cap −/+0.75 pt, margin +/−2 pts, build −/+$1.0M) and re-derive when Base changes, until you edit a cell, which then sticks.

Step 3

What it's worth

Complete your assumptions above to value the deal.
Enter MW, build cost, and rent (or contract value + term) to see what the deal is worth.
Estimated value created

Cost to build Value created = Stabilized value
Saves all your inputs into a link you can bookmark or share to reopen this exact scenario.

The number that decides everything: exit cap rate

The exit cap rate is the biggest driver of value. It reflects what return investors will accept for a fully leased data center. Drag the slider to test different scenarios: a lower rate is more optimistic (bull case), a higher rate is more conservative (bear case).

Bull Case · Base Case · Bear Case

The same deal under three coherent cases. Base is highlighted; each column header shows its cap rate · NOI margin · build cost.

Step 1

Add the company's colocation deals

Import a tracked deal or enter your own. We value each campus as if it is fully built and leased.
Step 2

Adjust the company's overall position

Here you can update net debt and shares. This step turns the total value of the deals into a value per share. Add a deal to the stack above first.
Other / unleased assets
$ · unleased campuses, land, other non-cash assets (cash is already in net debt)
Net debt
$ · total debt − cash
Shares outstanding
shares · preloaded count is as-reported basic. Add convertibles, warrants and RSUs under Additional dilutive
Additional dilutive optional
shares · warrants, RSUs
Scenario exit cap rates Applied to every deal in the stack. Bull and Bear derive from Base (−/+0.75 pt).
Bull
%
Base
%
Bear
%

Important: This calculation only includes the colocation deals you added above. It does not include mining operations, treasury, or any other campuses.

Step 3

Stabilized NAV per share

See the estimated value per share based on the deals and company position you entered above. Add a deal to the stack above to build the NAV.
Add at least one deal with a rent, margin, and cap rate to build the NAV bridge. Add shares outstanding to get NAV per share.
Stabilized NAV per share: Base case

This shows what the company could be worth per share if all the deals you added are fully built and leased. It does not account for construction delays or risks. Not a price target, and not today's book value.

We calculate the estimated market value of each campus once it is fully leased, then subtract any remaining construction costs.

NAV per share

We show three scenarios based on the different exit cap rates you set earlier.

This is an educational modeling tool, not investment advice or a recommendation. It simplifies deliberately: no maintenance capex, taxes, financing costs, or construction-period timing; single-deal debt attribution is not modeled. Location, power availability, tenant credit, and interest rates all matter and are not captured here. Disclosed figures come from the linked filings as of their dates; update assumptions when companies file new information. Know a deal we should track? Suggest it.

How it works & FAQ

Data center colocation deals are real-estate development deals, and this simulator values them the way developers do. The operator is a landlord: it builds the shell, the power infrastructure, and the cooling, then leases the finished capacity to a tenant who brings its own hardware. Two numbers drive everything. Yield on cost is the stabilized annual profit of the finished campus divided by what it cost to build. The exit cap rate is the yield a buyer would accept to own that same income stream once it is de-risked and cash-flowing. Because a signed, investment-grade lease is worth more finished than it costs to build, the cap rate sits well below the yield on cost, and that gap is where the value gets created. Use it as a colocation pricing calculator for a single lease, or stack deals into a company-level NAV per share.

Take TeraWulf's Anthropic lease: roughly $19B of contracted revenue over 20 years across 401 MW. Backing the escalator out gives about $725M of Year-1 rent (a flat ÷20 would overstate it), an 88% margin puts stabilized NOI near $638M, and against a ~$5.0B build cost that is a 12.7% yield on cost. Value the same NOI at a 6.5% cap rate and the stabilized campus is worth ~$9.8B, about $4.8B more than it cost to build, from one lease.

The model applies the cap rate to Year-1 stabilized NOI (the standard convention); the escalator affects only the rent conversion. It deliberately ignores financing, taxes, maintenance capex, and construction timing. It is an educational single-deal model, not a price target.

The Company NAV view stacks several deals into a net asset value per share, a sum-of-the-parts valuation. Each campus is valued at its stabilized asset value, then reduced by the remaining capex still to spend, never total build cost, because money already spent is sunk and already sits in net debt, so counting it twice would understate the company. Summing those net contributions, adding other or unleased assets, and subtracting net debt gives the company NAV; dividing by diluted shares gives NAV per share. It is a forward, stabilized figure: every deal valued as if delivered and fully ramped, with no discount for construction time or execution risk, so it deliberately shows no live share price and no implied upside. That comparison is yours to make.

The yield-on-cost vs exit-cap-rate framing this simulator is built on was popularized by @accounting_ds, whose breakdown of the TeraWulf × Anthropic lease inspired the model.

The capitalization rate is the annual net operating income of a stabilized property divided by its value: the yield a buyer accepts to own finished, leased, cash-flowing real estate. Lower cap rate = higher valuation for the same income. Prime single-tenant hyperscale campuses with investment-grade tenants trade at lower cap rates than riskier assets.

Stabilized annual NOI divided by total development cost. It measures what return the developer created by building rather than buying. When yield on cost exceeds the market cap rate, development creates value; the spread between the two is the developer's profit margin on the asset.

Because the entire stabilized valuation is NOI ÷ cap rate, small changes compound: on the TeraWulf reference deal every 0.5 pt of cap-rate compression adds roughly $700–970M of value. Nobody knows the right number in advance, and that is why the simulator makes you test a range instead of trusting a point estimate.

From Ziven's hyperscaler-deals dataset: every tracked deal is verified against a primary source (SEC filing or company press release) and linked. The import fills only what the company actually disclosed. Blanks mean the company didn't say, and that's information too. The dataset never invents a number.

AI cloud providers own the GPUs. Accelerators depreciate over roughly 3–5 years, so their cash flow is consumed by hardware refresh cycles, with economics closer to equipment leasing than to real estate. In colocation the chips sit on the tenant's balance sheet while the landlord's asset (shell, power, cooling) lasts decades. Same headline MW, completely different business.

EBITDA suits asset-light businesses like software. A colocation operator is the opposite: it is a landlord that spends heavily to build the campus and borrows to fund it. EBITDA hides the two costs that define these companies: the depreciation of what they built and the debt they raised to build it. That flatters the picture. These plays behave like REITs, so we value them the way REIT investors do: on stabilized cash flow (closer to AFFO than to EBITDA, because the landlord's maintenance capex is small once the shell, power and cooling are in) capitalized at a market cap rate. That is the yield-on-cost versus cap-rate framing this tool is built on.

A forward, stabilized NAV: every deal valued as if delivered and fully ramped, with no discount for construction time or execution risk. Not a price target, and not today's book value.

Retail colocation is typically quoted per rack or per kilowatt per month, often with power and cooling bundled in. The wholesale and hyperscale leases this simulator models are disclosed as a total contract value over a multi-year term for a stated number of megawatts. The simulator backs Year-1 rent out of those terms, and analysts compare deals in dollars per MW per year. A retail per-kW quote isn't directly comparable: it bundles services that a wholesale triple-net lease leaves to the tenant.
Value created
—