Methodology
The proposal you receive is generated by a deterministic economics engine. Every input is auditable. Two price-path scenarios sit above an empirical break-even floor — three numbers from one model, not three narratives.
Anchors
Break-even
Floor
Empirical lower anchor
The monthly BTC price at which production exactly clears electricity. Not a price forecast — a curve derived from your contracted hashrate, electricity at supplier cost, the halving schedule, and the projected network. Below this line you accumulate zero BTC.
House view
—
Committed forecast
One full halving cycle plays out, consistent with prior patterns. The number we ourselves are positioned around.
Bull
—
Institutional adoption
Institutional and sovereign adoption continues to compound through the cycle. We do not bet on this — we report what the model produces if it happens.
Break-even is computed, not set. The two terminal prices are set by Montblanc and adjusted as our outlook changes. They are not forecasts — they are the upper anchors the engine projects toward.
The model
01
A fixed quantity of contracted hashrate (TH/s), operated through the next halving in our facilities. You do not buy BTC and you do not buy the hardware — the hardware sits with Montblanc, operated on your behalf. What you buy is a contractual right to the BTC the contracted hashrate produces over the term, net of electricity.
02
Real mining against the live Bitcoin network. Each block pays a fixed subsidy and the fees in that block. Our share of every block is proportional to the network share your hashrate represents at the time. The mandate ends at the next halving (—) — the post-halving subsidy cut never falls inside the integration window, so the break-even curve we publish is the one you pay against.
03
The engine sweeps BTC/USD from today's spot to the scenario terminal price linearly in log-price over the mandate length (sign-date to halving). Two runs — house view and bull — produce two projected outcomes alongside the empirical break-even floor. There is no toggle; there is no curve-fitting.
04
Bitcoin's network hashrate grows continuously. Your share of every block — and therefore the BTC you receive — depends on how much hashrate is online globally at the time. The engine projects forward growth from the trailing four-quarter mean of net hashrate growth from non-shock quarters.
The April 2028 halving is modeled explicitly: empirical evidence from prior halvings shows a 3–6 month hashrate growth pause as marginal miners are forced offline. We freeze net growth for three months at the halving and resume from the post-halving level.
Confidence bands of ±1σ reflect the historical variance of non-shock growth. We do not extrapolate manufacturer roadmaps or hashrate-price elasticity — those models compound forecasting error across a multi-year horizon. The realized-Δ method has zero parameters to tune.
Honest limitation: public hardware shipment data covers only ~17% of supply (Canaan's NASDAQ filings). Bitmain (~82%) and MicroBT (~13%) are private and unobservable. We use realized network growth, not shipment forecasts, precisely because the shipment data does not exist.
The full method is auditable in the admin layer; the projection is deterministic from the underlying data.
05
The client receives 100% of the BTC produced by their contracted hashrate, less electricity at supplier cost. Montblanc's economics are operational — a maintained spare-unit pool, parts inventory, monitoring stack, software optimisation, and at-scale facility operations that produce more BTC per unit of capex than a single client could on their own. Production-cost economics for the client is the consequence of that operational efficiency, not a fee on the client's production.
06
Electricity is a real cost, not a margin lever. At — all-in, the engine projects the term electricity in BTC-equivalent at the integrated price path so you see term electricity in BTC alongside your net. In operation, electricity is invoiced monthly in fiat at supplier cost — no markup.
07
Every quote ships with a snapshot — the exact inputs the engine used. Re-running the snapshot produces the same numbers, byte-for-byte. Nothing in the engine reads market sentiment; the only knobs are the ones on this page.
08
The mandate is a service contract — the capex committed at sign is a prepaid business service expense, not a depreciable asset on your balance sheet (you don't own the hardware, Montblanc does). Electricity is operating expense; Bitcoin produced is business income at fair value on day of receipt. How each line is recognised — whether the prepaid service is amortised over the term (accrual basis) or deductible in the year of payment (cash basis), revenue timing, VAT treatment, holding-period classification — is determined by your accountant against the rules of your jurisdiction. We provide signed daily production records, monthly statements, on-chain settlement transaction IDs, the mandate contract, and the documentation your accountant needs. We do not provide tax advice. Treatment varies materially across France, Germany, Italy, the UK, Switzerland, Ireland, and the UAE; consult a qualified advisor before deploying capital.
What we do not promise
The numbers are projections, not forecasts. We do not know what bitcoin will trade at by the next halving. We model three plausible terminals and integrate the model against each.
Bitcoin is volatile. Network hashrate is competitive. Equipment fails. Facilities lose power. The engine accounts for these where they are measurable; it does not eliminate them.
Past production does not indicate future production. Past price action does not indicate future price action. The model is a discipline for sizing risk, not a substitute for due diligence.