Hashprice: the one number that tells you what Bitcoin mining pays

Ask what a mining machine earns and you will get a different answer for every machine. Ask what a terahash earns and there is exactly one answer for the whole network, updated continuously, and it is the number professional miners actually watch. It is called hashprice, and once you know it, every "is this miner profitable" question becomes one multiplication.

Right now hashprice is roughly $0.039 per TH/s per day — or, in the unit that does not move every time Bitcoin twitches, about 49 sats per TH/s per day. This page explains where that number comes from, how to use it, and why the sats version is the one worth remembering. Figures checked against on-chain data in the first half of September 2026.

The formula, worked in full

Hashprice is total miner revenue divided by total network hashrate. Nothing more.

Divide: 452.8 BTC ÷ 922,000,000 TH/s = 0.000000491 BTC per TH/s per day — 49.1 sats. At $78,000 per bitcoin that is $0.0383; at $81,000 it is $0.0398. Luxor's Hashrate Index — the industry reference — quoted spot hashprice at $39.36 per PH/s per day at the end of August, which is the same number in bigger units and confirms the arithmetic within a couple of percent.

That is the entire mechanism. Four inputs: subsidy, fees, Bitcoin price, network hashrate. Everything that happens to mining economics happens through one of those four doors.

Dollars lie to you; sats do not

Quoted in dollars, hashprice bounces with every move in the Bitcoin price and tells you nothing about your position in bitcoin terms. Quoted in sats, it moves only when difficulty or fees change — which is the decay that actually matters to a miner.

The sats view is brutal and clarifying: 49 sats per TH/s per day, falling every time difficulty rises, halving outright in April 2028. A machine's lifetime output in sats is close to deterministic. Whether those sats were worth buying the machine for is a Bitcoin price question — a separate bet you are making whether you notice or not.

What hashprice has actually done

Riot Platforms publishes the average hashprice it realised each quarter in its SEC filings — about the cleanest public record there is:

Read that column and you understand the last year of mining: revenue per unit of hashrate fell by half from the Q3 2025 peak to the June 2026 trough, then recovered 43% over the summer as Bitcoin climbed back toward $80,000. Every mining headline in that period — capitulations, AI pivots, fire-sale hardware — is this one column wearing different clothes.

Turning hashprice into machine revenue

Multiply by the machine's hashrate. That is all a mining calculator does before subtracting electricity.

Gross, not net. Electricity comes out of that, and electricity is where the outcome is decided: an S21 XP burning 84.2 kWh a day costs $0.61 at our cheapest hosted rate and $6.74 at 8-cent retail hosting — the difference between keeping 94% of gross and keeping 37%. The profitability model does this for every machine we stock at whatever rate you give it.

One honest note about that model: while writing this page we cross-checked its Bitcoin revenue constant against the on-chain figure and found it running about a third high — a leftover from when hashprice was stronger. It now uses the network-derived rate on this page, which means the profit figures across our product pages just got more conservative. We would rather show you smaller true numbers than larger stale ones.

Energy hashprice: the professional's version

Large operators quote revenue per megawatt-hour consumed rather than per terahash, because power is what they actually buy. The conversion folds in machine efficiency, and it splits the market in two:

Same coin, same network, same day — and one fleet earns 2.7 times more per unit of electricity than the other. That table is why old hardware sells for $1 per terahash while new hardware sells for $18: the market is pricing megawatt-hours of margin, not terahashes. And it is why CoinShares estimates 15–20% of the global fleet is underwater at industrial rates of 6 cents and above.

Where hashprice goes from here

Nobody knows, but the four inputs let you reason about it honestly.

Fees (0.6% of revenue): negligible now; any recovery is upside. Difficulty: rises almost monotonically — CoinShares models the network near 2 ZH/s by March 2027, which alone would cut sats-hashprice by more than half from today's 49. The subsidy: halves in April 2028, mechanically halving hashprice overnight — the full arithmetic is in our halving guide. Bitcoin's price: the only input that can push the other way, and the one nobody can model.

The sober conclusion: sats-per-terahash falls by design, forever. Mining stays viable by three routes only — cheaper power, more efficient machines, higher Bitcoin prices. The first one is the only one you fully control, which is the argument for hosting over home power in one sentence.

If you want the same arithmetic pointed at a different question — how long a machine takes to accumulate a whole coin — our guide to mining one bitcoin runs it, and the answer is measured in decades. For what today's hashprice means for the machine you are actually considering, put your electricity rate into the profitability ranking and read the net column, not the gross one.

Sources: mempool.space API for subsidy, fee averages, blocks per day and network hashrate; Luxor Hashrate Index for spot and historical hashprice including the August 2026 roundup; Riot Platforms quarterly SEC filings for realised hashprice by quarter; CoinShares Bitcoin Mining Report Q1 2026 for fleet and difficulty projections. Data as of early September 2026. Hashprice moves continuously; the sats figure decays with difficulty and halves at each halving. Mining income is never guaranteed.