Hashprice in 2026: The One Number That Prices a Mining Day

On 17 August 2026, the last weekly hashprice reading published by Luxor’s Hashrate Index came in at $31.89 per petahash per day. Ten days later, with Bitcoin trading near $79,000 after its strongest week in more than two years, the same machines were earning closer to $38 for the same work. No operator swapped a chip, rewired a substation, or shipped a single new rig. Hashprice is the number that explains how a miner’s paycheck can jump almost a fifth while the hardware on the shelf stays exactly the same.

Hashprice is the expected daily revenue a Bitcoin miner earns for each unit of computing power aimed at the network, quoted in US dollars per petahash per second per day, usually shortened to USD/PH/day. Divide by 1,000 and you get dollars per terahash per day, the unit that maps onto a single machine’s spec sheet. It is the closest thing proof-of-work has to a spot price for its core product, which is not bitcoin exactly but the raw hashing that secures the ledger and mints new coins.

This piece explains what hashprice measures, the four moving parts packed into it, why it swings so violently, how it quietly became a market you can trade and hedge rather than just a metric you read, and where the number stops being useful. The August 2026 rally is a convenient live illustration, because it stress-tested every one of those ideas at once.

What hashprice actually measures

Start with the product a miner sells. A mining machine does one thing: it computes SHA-256 hashes as fast as it can, hunting for a value below the network’s target. Its output is measured in hashes per second, scaled up to terahashes (TH/s, a trillion) and petahashes (PH/s, a quadrillion). What that output is worth in dollars depends on factors the miner does not control: how much bitcoin the network pays per block, how many other machines are competing, and what a bitcoin sells for. Hashprice bundles all of that into a single figure so an operator can answer one question: if I point one petahash at Bitcoin for a day, how many dollars come back?

The term is not generic jargon. Luxor Technologies, the mining-software firm behind the Hashrate Index, coined the word in 2019 and still publishes the reference series (docs.luxor.tech). The name is deliberate: it treats raw hashrate as a commodity with a price, the way a driller thinks in dollars per barrel or a farmer in dollars per bushel. A miner does not really sell bitcoin; a miner sells hashing and gets paid in bitcoin, and hashprice is the exchange rate between the two.

Two flavors exist, and keeping them straight is the single most useful discipline in reading the metric. USD-denominated hashprice, the headline number, tells a miner how many dollars a petahash earns per day. BTC-denominated hashprice strips the price of bitcoin out entirely and reports how much actual bitcoin that same petahash mints. The two tell very different stories, a point we return to below.

The four inputs, and the formula that ties them together

Hashprice looks like one number but it is really four, multiplied and divided into a single output. The revenue-side formula is straightforward: hashprice (USD/PH/day) equals (blocks per day x (block subsidy + average fees) x BTC price) divided by network hashrate.

Work through each term. Bitcoin targets roughly 144 blocks a day, one every ten minutes. The block subsidy has been 3.125 BTC since the April 2024 halving and stays there until the next halving, expected in spring 2028, when it drops to 1.5625. Average fees are whatever the transaction market adds on top, which in late August 2026 was under 1% of the reward. BTC price is the spot dollar value. Network hashrate is the total computing power competing for those blocks, currently a little over 900 exahashes per second.

Luxor’s published series expresses the same idea through difficulty rather than hashrate (hashprice equals (subsidy + fees) divided by difficulty, times the bitcoin price and a constant, with fees smoothed over a 144-block average). The two are algebraically the same, because difficulty and hashrate are locked together by the protocol’s ten-minute target. Whichever form you use, the intuition holds: three of the four inputs push hashprice up (a higher subsidy, higher fees, and a higher price all help), and one pushes it down (more competing hashrate divides the same pie into thinner slices).

A quick worked example makes it concrete. Take the network on 26 August 2026: 3.125 BTC subsidy, roughly 144 blocks, bitcoin at about $78,700 (Fortune), and hashrate near 920 EH/s, which is 920,000 PH/s. Multiply 144 by 3.125 by $78,700 and you get about $35.4 million of subsidy paid to miners in a day. Divide by 920,000 PH/s and the subsidy alone is worth about $38.5 per petahash per day, with the sub-1% fee layer adding a little more on top. That is a first-principles derivation, not a published quote, but it lines up with the roughly $38 to $39 that live trackers reported once bitcoin cleared $77,000 (BlockOps, Yahoo Finance).

Turning a terahash into a dollar

Because a single machine is rated in terahashes, the number that matters on a workbench is hashprice divided by 1,000. At roughly $38.5 per petahash per day, a miner earns about $0.0385 per terahash per day. An Antminer S21 XP, rated at 270 TH/s (Bitmain), therefore grosses around $10.40 a day before it pays for a single kilowatt-hour. Whether that $10.40 is a good business depends entirely on the power bill, which is the subject of a later section. First, the snapshot that frames the rest of this piece.

Two things jump out. First, the official weekly print and the real-time number were almost $6 apart, purely because bitcoin moved between the Monday the report was struck and the rally that followed. Second, the six-month forward already sits below the squeeze-era print, a signal we unpack when we get to the market for hashprice.

Why hashprice is the most volatile number in mining

Hashprice does not drift; it lurches. In the summer of 2025 it traded above $55 per petahash per day. By the end of June 2026 it had touched roughly $27, one of the lowest daily readings in five years, before the August rally dragged it back toward $38. That is a peak-to-trough drop of more than half, then a snap recovery of more than 40%, inside about a year.

The reason is that hashprice compounds two of the most volatile variables in the asset class. Bitcoin’s price is the obvious one. The quieter one is hashrate, which grows almost monotonically as manufacturers ship more efficient machines and operators plug them in. Every new terahash that joins the network competes with all the others for the same fixed 144 blocks a day, so rising hashrate mechanically dilutes everyone’s slice even when the price is flat. Through 2024 and 2025 hashrate roughly doubled, peaking near 1.1 zettahashes per second (about 1,100 EH/s) in October 2025 as bitcoin set its all-time high near $126,080. The subsidy, meanwhile, had just been cut in half in April 2024. Falling reward, surging competition, and a price that eventually rolled over from its record combined into the long grind lower that miners spent much of 2026 calling a capitulation.

This is why hashprice, not the bitcoin price, is the number miners actually watch. A 20% move in bitcoin is a headline; the same move in hashprice is the difference between a fleet that funds its own electricity and one that burns cash every hour it stays powered on. And because the price input is driven as much by macro forces as by anything on-chain, a single Federal Reserve decision or liquidity shift can reprice every rig in the world overnight.

The difficulty thermostat, mining’s self-correcting floor

The counterweight to hashprice’s volatility is one of Satoshi Nakamoto’s more elegant design choices: difficulty. Every 2,016 blocks, about every two weeks, the network recalibrates how hard it is to find a block so that blocks keep arriving roughly every ten minutes regardless of how much hashrate is online. When machines switch off, difficulty falls at the next retarget, and the miners still running capture a larger share of the same reward. Difficulty is a thermostat that quietly redistributes revenue back toward the survivors.

2026 has been a live demonstration. As hashprice sank below the cash cost of older fleets, operators unplugged their least efficient machines. By late August, difficulty had fallen to 125.81 trillion, just 0.7% above its June low and roughly 18% below the November 2025 peak, with an estimated 150 EH/s of capacity sitting idle (GNcrypto). It is on track for only the second year-over-year decline in Bitcoin’s history, a genuinely rare event in a network whose difficulty almost always rises.

Nikolaos Panigirtzoglou, the JPMorgan managing director who tracks mining economics, describes the loop plainly: “When bitcoin trades below its production cost, higher-cost miners power down, the hashrate declines, and difficulty adjusts lower.” His team pegged the all-in production cost of a marginal miner near $78,000 through much of 2026 (TFTC), which is why bitcoin spent months trading below the level at which the average public miner breaks even on a fully loaded basis. The August rally to $79,000 was the first time in roughly five months that the price closed the gap.

The thermostat cuts both ways, though. It cushions the downside, but it also competes away the upside. When a price rally lifts hashprice, idle machines switch back on and difficulty climbs at the next retarget, clawing back part of the windfall. That mean-reverting pull is exactly what the forward market was pricing in late August, as we will see.

Hashprice versus the bitcoin price, why they are not the same trade

It is tempting to treat hashprice as a leveraged bet on bitcoin. It is not, quite. The cleanest way to see the difference is to look at the BTC-denominated version of the metric.

In late August 2026, USD hashprice jumped from about $31.89 to roughly $38 because bitcoin rallied. But BTC-denominated hashprice, the amount of actual bitcoin a petahash mints per day, barely moved, holding near 0.00049 BTC. That tells you the rally was a price event, not a network event. Nothing about the competitive split of the reward across the world’s hashrate had changed; the same slice of bitcoin was simply worth more dollars.

The distinction matters because the two numbers answer different questions. A miner paying dollar-denominated power bills cares about USD hashprice: it decides whether the lights stay on. A long-term holder deciding whether to mine coins or simply buy them cares about BTC hashprice: it measures how efficiently hashing converts into satoshis over time, independent of price. BTC-denominated hashprice only improves when the subsidy rises (it never does between halvings), when fees spike, or when competing hashrate falls. Over any long horizon it grinds lower, halving by halving, which is the mathematical reason mining is a treadmill: you must keep getting more efficient just to mint the same bitcoin.

From revenue to profit, where hashprice meets the power bill

Hashprice is a revenue number. It says nothing about cost, and cost is where miners live or die. The bridge between the two is simple arithmetic. A machine’s electricity draw, set by its efficiency in joules per terahash (J/TH), determines how much power it burns to produce its hashrate. Set the revenue a machine earns equal to the power it consumes and you get the break-even electricity price: the dollars per kilowatt-hour at which the machine exactly covers its energy cost and not a cent more.

The formula miners keep on a sticky note is: break-even $/kWh equals hashprice divided by (24 times J/TH). Plug in the numbers and the fleet sorts itself into winners and casualties. The table below runs the current-generation and legacy machines at both the squeeze-era print ($31.89) and the post-rally level (about $38.33 reported by live trackers), using the standard published efficiency specs.

Read the table as a survival map. An operator buying power at $0.05 per kilowatt-hour runs every machine here at a profit in either column. One paying $0.07 was underwater on everything but the newest hardware during the squeeze, and comfortably positive across most of the fleet after the rally. A miner stuck at $0.10 kept only its S21-class rigs above water even post-rally, and an old S19j Pro at that power price never made sense at all. This is why the same $6 move in hashprice can be irrelevant to one operator and existential to another: it depends entirely on where each sits relative to its break-even line. It is also why the fine print of a miner’s power contract, and the after-tax treatment of what it mints, matters as much as the hashprice headline (a wrinkle for any US operator, since the tax bill on mined coins is one no broker files for you).

The rally illustrates a second, less obvious point: operating leverage. Because power cost is fixed while revenue is variable, a modest rise in hashprice can multiply a machine’s profit. One live analysis found that a 39% revenue increase turned into roughly a sevenfold profit increase on an S21 XP running at $0.08 power, precisely because the electricity bill swallowed nearly all of the pre-rally revenue (BlockOps). The leverage is largest for the least efficient machines, the ones closest to their break-even line, which is why marginal fleets swing from cash-negative to cash-positive on relatively small hashprice moves.

Hashprice became a market, not just a metric

For most of its life, hashprice was something miners read off a chart and worried about. Over the past two years it has quietly become something they can buy and sell. That shift is the most important development in mining finance, and it is where a metric turns into an instrument.

The logic is the same that gave farmers grain futures. A miner’s revenue is a stream of uncertain hashprice, and the miner would often rather lock in a known number than gamble on the spot market. Luxor now runs a hashprice forward market with contracts stretching up to twelve months, settling against the realized hashprice index (Luxor). On the exchange-traded side, Bitnomial lists Hashrate Futures under the ticker HUP, in one-petahash monthly contracts regulated by the Commodity Futures Trading Commission, giving miners and speculators a standardized, cleared way to take a view (Bitnomial). A miner short the forward is effectively pre-selling next quarter’s production at today’s price; a counterparty long the forward is buying exposure to mining revenue without ever racking a machine.

The shape of that forward curve is itself a forecast. In late August 2026, with spot hashprice near $38, the six-month forward traded around $30.67 per petahash per day, roughly 20% below spot, and contracts held above $35 only as far out as January 2027 before sloping down (Yahoo Finance). A downward-sloping, backwardated curve is the market’s blunt statement that it does not believe the rally’s hashprice will last: it expects difficulty to climb and the number to compress. Luxor’s July 2026 Lookback put the implied cost of capital baked into those forwards at roughly 6% to 13% annualized (Hashrate Index), a rough read on what miners will pay for revenue certainty.

This is also where hashprice stops being a Bitcoin-only story and starts resembling the plumbing of any commodity market: a physical producer, a financial hedge, a forward curve that encodes expectations, and a cost of carry. Bitcoin is not the only network that has to decide, block by block, who gets paid for securing it; other systems meter and price that work through entirely different mechanisms, as Bittensor’s Yuma Consensus shows in the staking world.

How a miner actually hedges hashprice

The mechanics are worth spelling out, because hedging hashprice sounds abstract until you follow the cash. Suppose a mid-size operator runs 5 exahashes (5,000 PH/s) and knows its all-in cost. At spot hashprice of $38, its fleet grosses about $190,000 a day; at $30 it grosses $150,000, and its margin evaporates. To protect its budget, the operator sells forward contracts covering, say, half its production at $34 per petahash per day for the next two quarters.

If hashprice then falls to $28, the operator’s physical revenue drops, but the short forward pays out the difference between $34 and the realized index, cushioning the blow. If hashprice instead rips to $45, the operator gives up part of that upside on the hedged half, paying the difference on the forward. Either way, the hedged production earns a known $34 regardless of what spot does. That is the entire point: a miner trades away the tails, in both directions, in exchange for a revenue number it can take to a lender or a landlord.

Contracts settle two ways. Financially settled instruments, the forwards and the Bitnomial futures, pay cash against the index and never involve delivering hashing. Physically settled or hosting-style deals, by contrast, sell the actual output of specific machines. The financial versions are what turned hashprice into a number a treasury desk can trade without owning a single rig, and they are why hashprice increasingly shows up in miners’ investor decks as a hedged, forecastable line rather than a spot lottery. The August rally was a small stress test of the discipline: operators who had hedged at squeeze-era levels underperformed the unhedged on the way up, which is the price of insurance and exactly what they signed up for.

The fee question and Bitcoin’s long-term security bill

One term in the hashprice formula is small today and enormous over time: transaction fees. In late August 2026, fees ran under 1% of the block reward, a rounding error next to the 3.125 BTC subsidy. For now, hashprice is almost entirely a subsidy story.

That will not always be true, and it cannot be. Every halving cuts the subsidy in half. After the spring 2028 halving the reward drops to 1.5625 BTC, and at a constant price that roughly halves the dollar value of the block reward and, with it, hashprice. The 2032 halving does it again. Over the long run, the only thing that can replace a shrinking subsidy is fee revenue, and fees are the one input miners and users cannot manufacture at will. They spike when blockspace is in demand: the launch of the Runes protocol in April 2024 briefly sent fees to record highs, with block 840,000 collecting 37.67 BTC in fees alone, the most expensive block ever at the time (CoinDesk). But those episodes fade, and fees have spent most of the time since well under 1% of the reward.

This is why hashprice times hashrate, the total dollars flowing to all miners in a year, doubles as Bitcoin’s security budget: it is roughly what an attacker would have to outspend to threaten the chain. At current levels that budget runs on the order of $13 billion a year, almost all of it subsidy. Whether fees can grow to carry that budget as the subsidy fades is one of the genuinely open questions in Bitcoin’s design, and it is bound up with whether new blockspace demand becomes durable rather than episodic, including the emerging world of Bitcoin rollups built on Taproot, BitVM and Citrea. Hashprice is where that abstract debate shows up as a concrete number on a miner’s dashboard.

When hashprice is not enough, the AI opportunity cost

The most consequential number in mining in 2026 is not hashprice itself but what hashprice is measured against. A megawatt of power and a data center to house it can be pointed at Bitcoin, or it can be leased to an artificial-intelligence tenant hungry for compute. That comparison now governs capital allocation across the sector.

The gap is stark when you convert hashprice into revenue per megawatt-hour, the unit a power-hungry business actually plans around. At the 17 August print, an efficient sub-14 J/TH fleet earned about $107 per megawatt-hour mining bitcoin; an inefficient 25 to 38 J/TH fleet earned roughly $41 (Hashrate Index). The rally lifted those figures by about a fifth. But signed AI and high-performance-computing leases can pay multiples of even the efficient tier, on fixed multi-year terms that do not swing with a difficulty retarget.

The right-hand column is a linear scaling of the cited 17 August figures to the post-rally hashprice, shown to make the direction clear, not a published quote. Even at the top of it, the comparison to AI is why the biggest listed miners are converting sites as fast as they can pour concrete.

The executives are blunt about it. John Todaro of Needham puts the math simply: “The revenue per megawatt and EBITDA margins are far higher for HPC and AI colocation than for mining.” MARA chief executive Fred Thiel frames the endgame in terms of power itself: mining, he says, “is a zero-sum game. As more people add capacity, it gets harder for everybody else. Margins compress, and the floor is your energy cost,” and his prediction is that “by 2028, you’ll either be a power generator, be owned by one, or be partnered with one” (CoinGeek). James Butterfill, head of research at CoinShares, estimates some listed miners could draw as much as 70% of revenue from AI by the end of 2026, up from roughly 30%, even as he calls a bitcoin recovery toward $100,000 “not an unrealistic assumption” that would lift hashprice back toward $37 (The Block).

For miners, then, hashprice is increasingly a hurdle rate. If a megawatt cannot earn more mining than it would earn leased to an AI tenant, the rational move is to lease it, and the fleet’s hashrate quietly leaves the network. The economics of that rival demand for compute, and how it is priced, is its own fast-moving story, one where decentralized inference faces its own commodity trap.

What hashprice cannot tell you

For all its usefulness, hashprice is an average, and averages hide the things that decide individual outcomes. A short list of what the headline number leaves out:

• Pool luck and payout scheme. Hashprice assumes a miner earns its exact statistical share of blocks. Real payouts depend on the pool’s model (FPPS, PPLNS, and their variants) and short-run variance; a small miner can run well above or below the theoretical number for weeks.
• The power contract. Hashprice is gross revenue. Two operators with identical machines and wildly different electricity deals (curtailment credits, demand-response payments, behind-the-meter generation) run completely different businesses at the same hashprice.
• Depreciation and hardware cost. The break-even math above covers electricity only. It ignores the capital sunk into the machines, which is why a fleet can be cash-positive and still lose money on a fully loaded, GAAP basis. Public miners routinely report both a cash cost to mine a bitcoin and a much higher all-in cost.
• Uptime and infrastructure. Curtailment for grid programs, maintenance, and heat all cut into the hours a machine actually runs. Hashprice assumes 24 hours a day; reality rarely does.
• Non-comparability across companies. Because each miner defines cost differently, the same hashprice supports very different reported margins.

That last point is worth a number. CoinShares pegged the weighted-average cash cost to mine a bitcoin near $80,000 in recent quarters, with individual operators ranging from the low $30,000s to figures distorted well past $100,000 by depreciation and accounting choices (CoinShares). In other words, hashprice tells you what a petahash earns on average, not what a specific miner keeps. It is the revenue line, not the bottom line, and treating it as the latter is the most common mistake in reading mining stocks.

Who regulates a hashprice, the SEC and CFTC boundary

Because hashprice now underlies tradeable contracts, it sits across a regulatory line that US crypto watchers know well. Two questions matter: is mining itself a security, and are hashprice derivatives commodities?

On the first, the Securities and Exchange Commission’s Division of Corporation Finance settled the near-term debate in March 2025, stating that certain proof-of-work mining activities, including solo mining and mining pools, are not offers or sales of securities, on the reasoning that miners rely on their own computational effort rather than the managerial efforts of others (SEC). That keeps the act of mining, and by extension the revenue hashprice measures, outside the securities framework.

The derivatives are a different regulator’s turf. Bitnomial’s Hashrate Futures are listed as commodity contracts under the Commodity Futures Trading Commission, the same body that oversees bitcoin futures. The clean way to hold the two ideas together: spot mining is not a security, and a financial contract written on mining revenue is a regulated commodity derivative. For an operator, the practical upshot is that hedging hashprice is a CFTC-world activity, with the disclosure and clearing that implies, while the mining itself carries no securities-registration burden in the United States.

The outlook, what the number is telling miners now

Strip away the detail and hashprice in late August 2026 is telling a two-part story. The near-term part is relief: the rally to $79,000 pulled hashprice off five-year lows and back above the cash cost of most of the fleet, and the falling difficulty of a shaken-out network means the survivors keep more of it. For the first time in months, the median public miner was roughly at, rather than below, its all-in cost.

The forward-looking part is caution. The six-month curve sitting 20% below spot is the market betting that difficulty will climb, that idle machines will switch back on, and that the rally’s hashprice will not hold. The 2028 halving looms as a scheduled cut to the number, and the AI opportunity cost keeps rising underneath it. Hashprice, in other words, is simultaneously the best it has looked all year and priced to fade. That tension, a relief rally the market does not trust, is the honest state of Bitcoin mining economics as the calendar turns toward autumn, with a Federal Reserve meeting in September and the usual macro cross-currents still capable of moving the price input more than anything happening on the network itself, the kind of two-clock countdown from Jackson Hole to September that miners now watch as closely as their own difficulty charts.

Frequently Asked Questions

What is hashprice in Bitcoin mining?

Hashprice is the expected daily revenue a miner earns for each unit of hashing power, quoted in US dollars per petahash per second per day (USD/PH/day). It bundles the block subsidy, transaction fees, the bitcoin price, and total network hashrate into one figure so an operator can see what a petahash of computing power is worth in dollars each day. Dividing by 1,000 gives the per-terahash figure that maps onto a single machine’s spec sheet.

How is hashprice calculated?

The revenue-side formula is (blocks per day times (block subsidy plus average fees) times the bitcoin price) divided by network hashrate. Bitcoin produces about 144 blocks a day at a 3.125 BTC subsidy, so most of the number comes from subsidy times price divided by the hashrate competing for it. Luxor’s Hashrate Index, which coined the term in 2019, publishes the same value using difficulty in place of hashrate, since the two are locked together by the protocol.

What is a good hashprice for miners in 2026?

There is no single answer, because it depends on a miner’s electricity cost and machine efficiency. As a rough guide, break-even electricity price equals hashprice divided by (24 times the machine’s J/TH). At the roughly $38 hashprice seen after the August 2026 rally, a modern Antminer S21 XP breaks even near $0.118 per kilowatt-hour, while an older S19j Pro breaks even near $0.054. Miners with power below about $0.05 per kilowatt-hour stayed profitable even through the 2026 squeeze; those paying $0.08 or more were often underwater on older machines.

Why did hashprice go up in August 2026?

Almost entirely because bitcoin’s price rose. Bitcoin rallied from around $64,000 to near $79,000 in its best week in more than two years, and since price is a direct multiplier in the hashprice formula, the number jumped from a published $31.89 to roughly $38 per petahash per day. The BTC-denominated version barely moved, confirming it was a price event rather than a change in mining competition. Falling network difficulty added a smaller boost.

Can you trade or hedge hashprice?

Yes. Luxor runs a hashprice forward market with contracts up to twelve months, and Bitnomial lists CFTC-regulated Hashrate Futures under the ticker HUP in one-petahash monthly contracts. A miner can sell forward to lock in a known revenue number, trading away both the downside and the upside on the hedged portion of its production. In late August 2026 the six-month forward traded around $30.67, about 20% below spot, signaling that the market expected the rally’s hashprice to fade.

Yuki Tanaka is a senior markets writer at HOGE Wire covering Bitcoin mining, energy, and the economics of proof-of-work.