Why capital binds
The binding constraint — the boundary this site runs into first — is capital, with a gap of 85 % ◌ derived. The runner-up is hardware at 54.7 % ◌ derived, a margin of 30.3 % ◌ derived.
The physics of this site clears; the financing does not. At the return its backers require, the income the site can earn does not cover what it costs to build and borrow for — so money, not power, heat or chips, is the boundary here.
Capital binds at 85 % ◌ derived. Levered NPV is -$54.8 M ◌ derived at a WACC of 10.6 % ◌ derived. The physical envelope clears; the financing does not.
In this market, demand is catching up with supply: utilisation pressure is building while the facility remains cash-generative. The constraint register files this state under the code R2 ◌ derived.
Constraints form a cascade: each solved constraint increases the load on the next boundary. If capital were cured — more of it bought, built or approved — the demand it now holds back would flow through to the next limit in this site's ranking, and hardware would become the binding constraint. Curing a constraint moves the boundary; it does not remove it.
Sweat the current fleetRefresh each cycleRefresh blocked
Counterfactual, sweat versus refresh: at 50 kW ◌ derived design density no next chip generation is supportable — the nearest higher generation enters at 120 kW ◌ derived per rack. The refresh path holds the current generation and matches the sweat path over the analysis horizon; the facility supports 1 ◌ derived generation classes over 5 years ◌ derived.
Whether that is acceptable is a judgment this model does not make.