Payback period is the time a miner takes to earn back what it cost. ROI (return on investment) is the total gain over a stated period divided by that cost. Both depend on revenue that changes every two weeks with network difficulty, so a calculation that freezes today's revenue will overstate the result. A useful estimate models revenue declining over time and shows how the answer moves when your assumptions do.
Key points
- Static payback = capital cost ÷ today's monthly margin. Treat it as the most optimistic case, not a forecast.
- Model a monthly decline in revenue and find two things: when cumulative margin reaches the capital cost, and when the machine stops covering its power bill.
- Electricity price changes the outcome more than any other input you control.
- Count everything in capital cost (freight, import charges, infrastructure) and everything in operating cost (all-in power rate, hosting, repairs).
- Nobody can promise a payback period. Test a range of assumptions with the calculator.
Payback period versus ROI
The two terms answer different questions.
Payback period asks when you get your money back. It is the number of months until cumulative operating margin (revenue minus operating costs) equals the capital cost. It ignores everything that happens afterwards.
ROI asks how much you gained relative to what you spent, over a period you must state:
ROI = (cumulative margin + resale value − capital cost) ÷ capital cost
A machine can have a short payback and a modest ROI if it becomes unprofitable soon after paying back, or never reach payback from margin alone and still end slightly ahead once sold. Always quote both, with the period.
Why static payback calculations mislead
The static calculation divides the price of the machine by today's daily profit. It assumes that profit continues unchanged, which has three problems.
- Difficulty growth. Revenue per TH/s is inversely proportional to difficulty. As more efficient hardware joins the network, difficulty rises and each existing machine earns fewer bitcoin per day.
- Halvings. The block subsidy is 3.125 BTC and halves every 210,000 blocks, next at block 1,050,000. A halving inside your planning period cuts the subsidy part of revenue in half from that block onward.
- Price volatility. Revenue in dollars moves with the Bitcoin price in both directions. A rising price can offset difficulty growth; a falling price compounds it.
The electricity bill does not shrink with revenue. As revenue falls toward a fixed power cost, the margin shrinks much faster than the revenue does, and at some point the machine no longer covers its own electricity.
A model with declining revenue
The simplest improvement is to let revenue fall by a fixed percentage each month. That percentage stands for the net effect of difficulty, price and fees. If difficulty grew 3% in a month while the price rose 1%, revenue would change by 1.01 ÷ 1.03 − 1 = −1.9%.
Small monthly rates compound into large annual ones:
| Monthly decline | Revenue remaining after 12 months | After 24 months |
|---|---|---|
| 1% | 88.6% | 78.6% |
| 2% | 78.5% | 61.6% |
| 3% | 69.4% | 48.1% |
The model used below:
- Machine: 200 TH/s at 3,500 W (illustrative).
- Capital cost: $2,400, all-in (illustrative).
- First-month revenue: $248.40. This is $8.28 per day after pool fees for 30 days, from the illustrative inputs in How Bitcoin Mining Profitability Works. It is not a current figure.
- Revenue in month m = $248.40 × (1 − r)^(m − 1), where r is the monthly decline.
- Electricity: 3.5 kW × 720 h = 2,520 kWh per 30-day month, at a constant price. No other operating costs.
- The machine runs continuously and is switched off for good in the first month its margin would be zero or negative.
- Horizon 60 months. No halving is modelled, and no resale value is included.
Sensitivity table
Months until cumulative margin reaches the $2,400 capital cost. "Not reached" means the machine was switched off, or the 60 months ended, before that point.
| Electricity price | Monthly power cost | Static payback | 0% decline | 1% decline | 2% decline | 3% decline |
|---|---|---|---|---|---|---|
| $0.03/kWh | $75.60 | 13.9 months | 14 | 16 | 18 | 23 |
| $0.04/kWh | $100.80 | 16.3 months | 17 | 19 | 25 | Not reached |
| $0.05/kWh | $126.00 | 19.6 months | 20 | 26 | Not reached | Not reached |
| $0.06/kWh | $151.20 | 24.7 months | 25 | Not reached | Not reached | Not reached |
| $0.08/kWh | $201.60 | 51.3 months | 52 | Not reached | Not reached | Not reached |
Static payback is $2,400 divided by the first-month margin, for example $2,400 ÷ ($248.40 − $100.80) = 16.3 months. The 0% column is the same result rounded up to a whole month.
Cumulative margin over the horizon, in the same model, shows how far each "not reached" case fell short:
| Electricity price | 0% decline | 1% decline | 2% decline | 3% decline |
|---|---|---|---|---|
| $0.03/kWh | $10,368 | $6,713 | $4,189 | $2,808 |
| $0.04/kWh | $8,856 | $5,201 | $2,880 | $1,936 |
| $0.05/kWh | $7,344 | $3,689 | $1,887 | $1,273 |
| $0.06/kWh | $5,832 | $2,252 | $1,145 | $776 |
| $0.08/kWh | $2,808 | $493 | $257 | $179 |
These figures are outputs of a simple model with illustrative inputs. They show sensitivity, not expected earnings.
What the tables show:
- The static figure hides most of the risk. At $0.06 per kWh the static payback of 24.7 months looks workable. With a 2% monthly decline, margin turns negative in month 26 after recovering $1,145, less than half the capital cost.
- Power price changes the shape of the result, not only its size. At $0.03 the machine pays back under every decline rate shown. At $0.08 it pays back under none of them unless revenue never falls.
- ROI collapses faster than payback suggests. At $0.04, 60-month ROI is ($8,856 − $2,400) ÷ $2,400 = 269% in the static case and ($2,880 − $2,400) ÷ $2,400 = 20% with a 2% decline, although payback moves only from 17 to 25 months.
- A halving would make every declining case worse. If one falls inside your horizon, model it as a step down in revenue in that month.
The model can also err in the pessimistic direction. Revenue does not fall smoothly, and a rising Bitcoin price can lift revenue per TH for months at a time. The point is to see the range, not to pick one row and believe it.
Residual value
A machine that is switched off is not worth zero. It can be sold to an operator with cheaper power, or kept for parts. Resale value belongs in ROI:
net result = cumulative margin + resale value − capital cost
Take the $0.06 per kWh, 1% decline case. Margin alone recovers $2,252 over the 50 months the machine runs, short of the $2,400 cost. If the machine then sold for an illustrative $400, the net result would be $2,252 + $400 − $2,400 = $252, an ROI of 10.5% over 50 months.
Resale value is itself uncertain. It falls as difficulty rises, and it falls fastest for less efficient machines, because fewer sites can still run them at a profit. Efficient machines generally hold value longer. See ASIC Miner Lifespan.
What to include in capex and opex
A payback figure is only as good as the costs behind it.
Capital cost (capex)
- Machine price, including the power supply and cables if sold separately.
- Freight, insurance, import duties, taxes and brokerage.
- Electrical work: circuits, PDUs, switchgear, transformers.
- Cooling: ventilation for air-cooled machines; pumps, dry coolers and piping for hydro; tanks and fluid for immersion.
- Racks, networking, installation labour and any hosting set-up fees or deposits.
Operating cost (opex)
- Electricity at your all-in rate, including demand charges and cooling overhead. See Electricity Cost and Bitcoin Mining.
- Hosting fees, if the machine is hosted.
- Pool fees, if not already deducted from revenue.
- Repairs, spare parts, replacement fans and labour.
- Downtime. A machine that runs 95% of the time earns 95% of the modelled revenue.
- Internet, insurance, rent and taxes where they apply.
Running your own estimate
- Get revenue per day for the machine at current network values from the calculator.
- Work out your all-in electricity price per kWh.
- Add up the full capital cost per machine.
- Choose at least three monthly decline rates and compute payback and the shut-off month for each.
- If a halving falls within your horizon, apply it.
- Add a conservative resale value and compute ROI over the period.
- Compare the result with the alternative use of the same money, including simply buying and holding bitcoin.
On the last point: at an illustrative price of $100,000, a $2,400 machine costs 0.024 BTC. The mining investment beats holding only if the bitcoin it mines, net of the bitcoin-equivalent cost of electricity, exceeds that amount.