Methodology

The question

A profitable US business has pre-tax net profit available to deploy toward Bitcoin. Two paths: buy miners and pay hosting, or pay tax first and DCA the net into spot BTC. The calc compares identical pre-tax dollars committed to each over a 4-year horizon (one halving cycle).

Why 4 years

One halving cycle. Mining decisions past 4 years are dominated by assumptions we can't credibly make: future ASIC efficiency, future hosting markets, future tax law. A typical operator runs hardware through the full cycle. Some hold longer based on their own economics — that's the operator's call to make with fresh data, not something we'll project.

The Mine path

Your pre-tax capital buys miners. Because ASIC equipment qualifies for 100% bonus depreciation under OBBBA 2025 (property placed in service after January 19, 2025), the entire purchase is deductible year one. That deduction refunds cash at your marginal rate, which buys more miners. Mathematically, CapEx amplifies by 1 / (1 − marginal_rate).

The fleet then runs for 48 months. Each month, mined BTC accumulates at (your_TH / network_TH) × block_subsidy × blocks_per_month × (1 + fee_share) × (1 − pool_fee) × (1 − host_profit_share).

Hosting bills are paid from ongoing business cash (default mode). Network hashrate grows per the difficulty growth preset (moderate default: 15% year one, decaying to 8% terminal). The 2028 halving inside the horizon cuts block subsidy from 3.125 to 1.5625 BTC at month 24.

At year 4, the fleet is sold at the resale-curve fraction (default: 15% of original, per Hashrate Index ASIC Price Index new-gen tier). The sale triggers §1245 recapture — resale proceeds are taxed as ordinary income at your marginal rate, up to prior deductions. The BTC stack is then assumed liquidated; gains above cost basis (sum of USD value of mined BTC at the time of each block reward) incur LTCG at 20% federal.

The Buy path

Your pre-tax capital is taxed at your effective rate. The net is deployed to BTC via the selected onramp (River, Swan, Bitcoin Well, etc.), DCA'd evenly over 12 months. This matches how a real treasury allocation works — not lump-sum at day one.

Fair parity: because the Mine path is committing ongoing operating cash to hosting, the Buy path receives matching ongoing cash as continued DCA across all 48 months. Both paths commit identical total pre-tax dollars — the only difference is the deployment mechanism.

At year 4, the stack is liquidated; gain above basis pays LTCG at 20% federal.

Sazmining profit-share model

Traditional hosts (Compass, Simple Mining) profit by marking up electricity and charging a higher all-in $/kWh rate. Sazmining's aligned-incentive model is different: they charge a lower service fee (approximately at-cost electricity) and take a percentage of mined BTC at the pool level before payout. Paraguay and Ethiopia sites take 15%; Norway takes 20%.

The calc applies the management fee as a BTC haircut on every mining payout, on top of the regular pool fee. This is not free: a 15% haircut on mined BTC is roughly equivalent to adding 1.5–2¢ to the effective $/kWh cost in sat-yield terms. The visible electricity rate is not the full story.

BTC price model

Default: Power Law (Santostasi). Price scales with (time since genesis)^5.8. Fifteen-plus years of BTC price action follow this trajectory closely. At current BTC age this implies roughly 25–35% forward CAGR over 4 years, decaying as the asset matures. The base case if BTC continues doing what it has always done.

Stress test: Doomer — flat at today's price for 4 years. This scenario has never happened — BTC has no zero-CAGR 4-year window in its history, including 2018 and 2022 bear-market entries. Shown because "what if the next cycle looks like no cycle has ever looked" is a legitimate downside question.

Bullish: Optimist — 40% CAGR. Below historical realized performance: 2017→21 was ~60%, 2020→24 was ~55%, 2021→25 was ~20%. This assumes halving-driven supply shock continues to meaningfully outpace demand growth.

User can also supply a custom CAGR.

Data sources

• BTC spot price: CoinGecko, cached 5 min at edge
• Network hashrate: mempool.space /api/v1/mining/hashrate, cached 1 hr
• Difficulty adjustment: mempool.space /api/v1/difficulty-adjustment
• ASIC prices: OneMiners, ASIC Marketplace, Compass Mining, Simple Mining (April 2026 retail)
• Hosting rates: public pricing pages — compassmining.io, simplemining.io/hosting, sazmining.com
• Tax law: OBBBA 2025 (100% bonus depreciation restored), IRC §1245 (recapture on equipment sale), IRC §1(h) (LTCG)

Limitations

This calc does not quantify: counterparty risk (host bankruptcy, equipment seizure), uptime variance below nameplate, firmware degradation, transmission losses, grid instability, regulatory changes, tariff risk on imported hardware, insurance cost, or the opportunity cost of management attention. These are real and can shift the answer materially.

It also assumes the hosting contract is honored at the stated rate for the full 4 years. Real-world contracts have termination clauses, rate-escalation clauses, and force majeure provisions — read yours before committing.

The math is honest. Reality is messier.