What mining profitability actually measures
Mining profitability is a rate, not a return. It answers one question: for every day this machine stays powered on, does the coin it earns sell for more than the electricity, pool fee and hosting it consumes? Hardware cost does not appear anywhere in that question — that is a separate calculation, and you should run it separately in the mining hardware ROI and payback calculator.
The revenue half rests on one idea: proof-of-work pays you in proportion to the work you contribute. If you supply 0.1% of the network's hashrate, you expect 0.1% of every block reward the network produces. You are not being paid for solving anything clever; you are being paid for the fraction of the total guessing you did. That makes expected revenue exactly linear in your hashrate and exactly inverse in the network's.
The cost half is far more predictable than the revenue half. A machine drawing 3,250 W consumes 78 kWh a day whatever the price of the coin does, and that bill arrives monthly. Revenue swings with price, difficulty and fee pressure; cost barely moves. That asymmetry is why experienced operators watch the break-even electricity price rather than today's profit — it is the number that changes slowly.
One caution about the word expected. Through a pool, your share arrives smoothly and the expected value is close to what you actually receive week to week. Mining solo, the same expected value is the mean of a very skewed distribution: most weeks you get nothing at all. The mining difficulty and block time calculator works out that distribution.
The formula, term by term
Revenue is built from four multiplications, each of which you can sanity-check on its own.
Your share of the network is H ÷ N. Both must be in the same unit. Almost every wrong answer people get from a mining calculator comes from this single division — a machine quoted in TH/s divided by a network quoted in EH/s, giving an answer a million times too large. The unit selectors on both fields exist for exactly this reason.
Blocks per day is 86,400 ÷ T. Bitcoin targets 600 seconds, so 144 blocks a day. This is the protocol's target, and difficulty retargeting is what holds the real average near it: Bitcoin adjusts every 2,016 blocks so that those blocks take about two weeks. Between retargets the real rate drifts above or below 144 whenever hashrate joins or leaves.
Coins per block is the subsidy R plus average fees F. The subsidy is fixed by consensus and halves every 210,000 blocks; fees are whatever the mempool will bear that day, and they are the volatile part. Use a multi-day average, because a single congested day is not a forecast.
Price and pool fee convert coins to cash. The pool fee is charged on your payout, so it scales with revenue rather than being a fixed cost — which means it never affects your break-even electricity price by as much as people expect.
On the cost side, W ÷ 1000 × 24 converts watts to kilowatt-hours per day, and multiplying by C gives the daily bill. Use the wall figure. The number printed on an ASIC's spec sheet is often measured at an ideal ambient temperature and a 240 V supply, and real installations commonly read higher once power-supply losses and fans are included.
Setting net profit to zero and solving for C gives the break-even electricity price; solving instead for P gives the break-even coin price. Those two rearrangements are the most useful thing on this page.
Worked example: one 110 TH/s machine against a 700 EH/s network
Take a single ASIC rated 110 TH/s drawing 3,250 W at the wall. The network is running 700 EH/s, the subsidy is 3.125 BTC, fees average 0.05 BTC per block, the price is $100,000, the pool charges 2%, and power costs $0.06/kWh.
- Put both hashrates in the same unit. 700 EH/s = 700,000,000 TH/s.
- Your share. 110 ÷ 700,000,000 = 1.5714286 × 10−7, which is 0.0000157% of the network.
- Blocks per day. 86,400 ÷ 600 = 144.
- Coins per block. 3.125 + 0.05 = 3.175 BTC.
- Coins per day. 1.5714286 × 10−7 × 144 × 3.175 = 0.00007184571 BTC.
- Gross revenue. 0.00007184571 × $100,000 = $7.1846 per day.
- Pool fee. $7.1846 × 0.02 = $0.1437.
- Electricity. 3,250 W ÷ 1000 × 24 = 78 kWh; 78 × $0.06 = $4.68.
- Net profit. $7.1846 − $0.1437 − $4.68 = $2.3609 per day.
That is $71.86 a month (× 30.4375) and $862.31 a year (× 365.25). Net margin is 2.3609 ÷ 7.1846 = 32.9%, and the all-in cash cost of the bitcoin you produce is ($0.1437 + $4.68) ÷ 0.00007184571 = $67,140 per BTC.
Now the two break-evens. Revenue after the pool fee is $7.0409 a day, spread over 78 kWh, so you break even at 7.0409 ÷ 78 = $0.0903 per kWh. At the same $0.06 rate, the price at which profit hits zero is $4.68 ÷ (0.00007184571 × 0.98) = $66,469 per BTC. Both figures come straight out of the same nine numbers you already entered — nothing else is hiding in them.
How to read the result
Read the break-even electricity price first. It compresses your machine's efficiency, the coin price and the network's size into one number you can compare directly against your utility bill. If your rate sits well under it, the machine survives a large fall in price or a large rise in difficulty before it has to be switched off. If your rate is just under it, you are one retarget away from running at a loss.
The distance between your rate and the break-even rate is your margin of safety, and it is worth stating as a ratio. In the worked example, $0.06 against a $0.0903 break-even means revenue could fall by a third before the machine stops covering its power. Operators who quote a single profit figure without that ratio are quoting the least stable number on the page.
Net margin is the second thing to read, and it behaves in a way that surprises people: it is not a property of the machine. Two identical machines at different electricity rates have different margins, and the same machine's margin swings from comfortable to negative over a single price cycle, because the cost side barely moves while the revenue side halves. Treat margin as a snapshot, not a characteristic.
Finally, compare the all-in cash cost per coin against the market price. If your cost per coin is below the price, you can sell production and stay cash-positive. If it is above, every coin you mine is worth less than the power that made it, and holding rather than selling is a bet on price, not a mining decision.
Nothing on this page includes hardware depreciation, which for a machine on a two-year replacement cycle can be the largest cost of all. Cash profit and economic profit are different questions.
Break-even electricity price by machine efficiency
| Efficiency (J/TH) | $0.04/TH/day | $0.05/TH/day | $0.06/TH/day | $0.08/TH/day |
|---|---|---|---|---|
| 15 | $0.111 | $0.139 | $0.167 | $0.222 |
| 20 | $0.083 | $0.104 | $0.125 | $0.167 |
| 25 | $0.067 | $0.083 | $0.100 | $0.133 |
| 30 | $0.056 | $0.069 | $0.083 | $0.111 |
| 35 | $0.048 | $0.060 | $0.071 | $0.095 |
| 45 | $0.037 | $0.046 | $0.056 | $0.074 |
| 60 | $0.028 | $0.035 | $0.042 | $0.056 |
Multiply the figure by (1 − pool fee) to include the pool. The worked example above is 3,250 W ÷ 110 TH/s = 29.5 J/TH at a hashprice of $0.0653/TH/day, which lands between the 25 and 30 J/TH rows — and after the 2% pool fee gives exactly the $0.0903/kWh computed there.
Difficulty moves against you by design
Every mining calculator, this one included, holds the network hashrate fixed. The protocol does not. Bitcoin retargets difficulty every 2,016 blocks — roughly every two weeks — so that blocks keep arriving near the 600-second target. When hashrate joins the network, difficulty rises and your share of each block falls by the same proportion. Your revenue per unit of hashrate is therefore inversely proportional to total network hashrate, and a single machine's output falls whenever the network grows, even though the machine itself has not changed.
Treat any figure here as a snapshot at today's difficulty. To carry an assumption about network growth into a multi-month projection, use the difficulty-growth input in the payback calculator.
Mistakes that make a profitability estimate wrong
- Mixing hashrate units. TH/s against EH/s is a factor of one million. Check that your share of the network is a plausible fraction before you believe anything downstream of it.
- Using spec-sheet power instead of wall power. Power-supply losses, fans and a warm room all raise the number your meter reads.
- Using the headline tariff rather than the delivered rate. Demand charges, transmission and taxes belong in the electricity price field; miners routinely underestimate it.
- Forgetting that difficulty ratchets. Today's coins-per-day is the best case for every future day unless the network shrinks.
- Assuming a single congested day's fees are typical. Fee revenue is the most volatile input on the page. Average it over at least a week.
- Ignoring downtime. A machine that runs 95% of the time earns 95% of this revenue but often pays close to full hosting.
- Confusing cash profit with return on capital. A rig can be cash-positive for two years and still never repay what you paid for it.
- Overlooking tax. In many jurisdictions mined coins are ordinary income at the fair market value on the day they are received, and a later sale is a separate taxable event.
Where this sits among the other numbers miners quote
The industry increasingly quotes hashprice — gross revenue per unit of hashrate per day, in $/TH/day — instead of coins per day. It is the same quantity with the machine divided out: hashprice equals your gross daily revenue divided by your hashrate, and it lets you compare two machines of different sizes at a glance. In the worked example, $7.1846 ÷ 110 TH/s = $0.0653/TH/day. Because hashprice already folds price, difficulty, subsidy and fees into one figure, a fall in it is the single cleanest signal that mining economics have tightened.
The mirror-image metric is efficiency in joules per terahash, which is just watts divided by TH/s. Efficiency and hashprice together determine break-even electricity price and nothing else does, which is why the reference table above needs only those two columns.
Two adjacent questions need different tools. To decide whether a machine repays its purchase price, you need payback and ROI, which depend on capital cost and a holding period — that is the payback calculator. To work out how long a given hashrate waits for a block, you need the Poisson model of block discovery rather than an average, which is the difficulty and block time calculator. And if your interest is in transaction costs rather than block rewards, the Ethereum gas fee calculator handles the other side of the ledger.
Finally, mining is an energy business wearing a computing costume. The same arithmetic that decides whether a rack of ASICs pays also decides whether a rack of servers pays; if you are comparing self-hosting against rented capacity, the cloud VM monthly cost calculator uses the same cost-per-hour reasoning.
Key terms
- Hashrate
- The rate at which a machine computes candidate block headers, in hashes per second. One TH/s is 1012 hashes per second.
- Difficulty
- A protocol-set number that scales how hard a valid block hash is to find. It retargets so that block intervals stay near the protocol's target.
- Block subsidy
- Newly created coins paid to whoever finds a block. Fixed by consensus; Bitcoin's halves every 210,000 blocks.
- Hashprice
- Gross mining revenue per unit of hashrate per day, usually quoted in $/TH/day. It combines coin price, difficulty, subsidy and fees into one number.
- Efficiency (J/TH)
- Wall power divided by hashrate. Numerically the same as watts per TH/s. Lower is better; it fixes how many kWh each terahash-day costs you.
- Pool fee
- The share a mining pool keeps from your payout in exchange for smoothing your income. Charged on revenue, so it scales rather than being fixed.