Hashprice explained: the number that tells you what a machine earns

If you follow mining for a week you will hear three numbers thrown around: the coin price, the difficulty and the hashrate of the network. Each of them tells you something. None of them tells you what your machines earn. Hashprice does, and it is the number to build a plan on.

What hashprice is

Hashprice is mining revenue per unit of computing power per day. It is usually quoted in dollars per terahash per day, sometimes in satoshis. If hashprice is 0.045 dollars per terahash per day, then a 200 TH/s machine grosses about nine dollars a day before electricity. That is the whole calculation. Multiply your fleet's terahash by hashprice and you have gross revenue.

The elegance of it is that it hides three moving parts behind one figure. Coin price, network difficulty and the fee share of block rewards all feed in. You do not have to track them separately to know what a day is worth.

Why it moves

Hashprice rises when the coin price rises, and falls when more computing power joins the network, because the same rewards are then split more ways. Those two forces frequently pull in opposite directions, which is why a rising coin price does not always mean a rising hashprice. A bull market attracts hardware, hardware raises difficulty, difficulty pushes hashprice back down. The lag between the two is where a lot of money is made and lost.

Transaction fees add a third layer. In quiet periods fees are a rounding error. In congested periods they have briefly been a large share of what miners collect. Fee spikes are real revenue, but they are not something to underwrite a purchase on.

Turning hashprice into your margin

Gross revenue is hashprice times terahash. Your cost is power draw times your electricity price times hours. Subtract, and the difference is what the machine actually contributes. Worked through for a single unit at 200 TH/s and 3,500 W on an eight cent power price: revenue at 0.045 hashprice is nine dollars a day, power is 3.5 kW times 24 hours times 0.08, which is 6.72. The machine contributes 2.28 dollars a day. On two hundred units that is 456 dollars a day, and the number is now big enough to plan against.

The same machine on a fifteen cent power price loses money at that hashprice. Nothing about the hardware changed. This is the point our note on the breakeven power price makes from the other direction: your electricity contract, not your hardware, is usually the binding constraint.

What hashprice does not tell you

• It says nothing about your uptime. A machine hashing 88 percent of the month earns 88 percent of the number.
• It says nothing about pool fees, which come off the top before you see anything.
• It ignores curtailment. If your contract requires you to shut down on peak days, your effective hashprice for the month is lower than the published one.
• It is a spot figure, not a forecast. Building a five year model on today's hashprice is not planning, it is hoping.

Planning with it honestly

Use a band, not a point. Run your numbers at today's hashprice, at half of it and at one and a half times it. If the machine still works at half, you have a purchase that survives a bad year. If it only works above today's figure, you are making a directional bet on the coin price and you should say so out loud, because that is a different decision with a different risk.

The other habit worth building is watching hashprice against hardware prices. When hashprice falls and machine prices have not caught up yet, buying is expensive in real terms. When hashprice recovers and hardware is still cheap, that window is short. Efficient machines survive low hashprice periods longest, which is the argument in efficiency over hashrate.