Hashprice in 2026: The Ceiling Is the Network, Not the Price

The number that recovered, and the number that keeps taking it back

Through the summer of 2026, Bitcoin miners lived on the thinnest revenue of the post-halving era. Hashprice, the single number that tells an operator what a unit of computing power earns in a day, slid to roughly $27.70 per PH/day at the end of June, one of the lowest daily readings in five years. By the first week of October it had clawed back to around $40 per PH/day, according to Hashrate Index, a recovery of nearly 50 percent, carried by one of Bitcoin’s strongest quarters in two years.

And yet the recovery answers to a governor that no central bank, no regulator, and no exchange controls. On September 19 the network’s difficulty adjustment jumped 4.16 percent, the largest single step up since June; on October 3 it held essentially flat at 132.72T (CoinWarz). Each of those moves does something subtle but decisive: it changes how much of any Bitcoin price gain actually reaches the people running the machines. The argument of this piece is simple. The real ceiling on a miner’s recovery in 2026 is not the Federal Reserve and not the SEC. It is the network itself.

To see why, you have to understand what hashprice is, where difficulty sits in its math, and why the two are locked in a feedback loop that competes away every windfall almost as fast as the price can create it.

What hashprice actually is

Hashprice is the expected revenue a miner earns per unit of hashrate per day. It is usually quoted in US dollars per petahash per second per day (written $/PH/day), or divided by 1,000 for terahash. Luxor’s Hashrate Index desk coined the term in 2019, and it has become the mining industry’s most-watched gauge because it folds four moving parts, the block subsidy, transaction fees, the Bitcoin price, and network difficulty, into one figure an operator can hold up against a power bill (Luxor documentation).

If you own a machine that produces 100 TH/s and hashprice is $40 per PH/day, that rig is expected to earn roughly $4.00 of Bitcoin a day before electricity (100 divided by 1,000, times $40). Scale the same math to a 1 EH/s fleet and you are looking at about $40,000 a day of gross revenue. Hashprice is deliberately agnostic about your costs; it measures the top line and leaves the expense side (power, hosting, depreciation) to you. That division of labor is exactly why it travels so well across wildly different operations. For a hobbyist pointing a single box at the network, the same number explains the brutal arithmetic of solo mining’s lottery odds.

Because it is expressed per unit of work, hashprice lets a hobbyist with one machine and a public company with 50 exahashes read the same signal. When analysts say mining revenue rose or fell, hashprice is almost always the number sitting underneath the headline.

The formula, and why difficulty sits in the denominator

The intuitive version is daily network revenue divided by daily network work. There are about 144 blocks a day. Each one pays a subsidy of 3.125 BTC (the level since the April 2024 halving) plus whatever transaction fees ride along. Multiply by the Bitcoin price and you have the dollars flowing to all miners in a day. Divide by total network hashrate and you have the dollars per unit of hashrate, which is hashprice.

Written compactly: hashprice is approximately (144 times (subsidy plus fees) times the BTC price) divided by network hashrate. Luxor’s canonical version puts difficulty in the denominator instead of hashrate, because the two are tied together: hashprice equals ((subsidy plus fees) divided by difficulty) times the BTC price, with fees smoothed over a 144-block window.

Try it with early-October numbers. 144 blocks times 3.125 BTC is 450 BTC of subsidy a day. At a Bitcoin price near $84,864 (CoinGecko), that is about $38.2 million. Divide by a network running near 950 EH/s (CoinWarz) and you land close to $40 per PH/day, which matches the quoted figure. Fees would add a fraction of a percent on top.

Now the structural point that this whole article turns on: the Bitcoin price is in the numerator, and difficulty (hashrate) is in the denominator. Price lifts hashprice directly. But difficulty pushes back directly too, in the opposite direction, and it is the only input that responds automatically to the others. The subsidy is fixed until 2028. Fees are tiny. The price is set by a global market no single miner controls. Difficulty is the one variable that moves because of what miners themselves do, which is precisely what makes it a ceiling rather than a tailwind.

Early October 2026, by the numbers

Here is where the master metric sits in the first week of October 2026.

Two things jump out of that table. The dollar figure has recovered hard, while the Bitcoin-denominated figure has barely budged. And difficulty, having fallen all summer, has climbed back to within about 15 percent of its November 2025 record. Both facts point to the same conclusion, which the next sections unpack in turn.

The recovery: from a five-year low to about $40

Hashprice’s 2026 has been a round trip. It opened the year near $37.60, ground down through a miserable first half to a late-June trough around $27.70, then recovered through the autumn as Bitcoin staged its best quarter since 2017. By early September the number had already risen 22 percent in a single month to $39.63, a jump reported by news.bitcoin.com as difficulty began to turn back up.

The driver was almost entirely price. Bitcoin climbed from the low $60,000s in midsummer to the mid $80,000s by early October, and because the price sits in hashprice’s numerator, the revenue line followed. CoinShares research chief James Butterfill had called the stretch one of the most challenging periods for miners since the halving; the rally finally gave the survivors some room to breathe. But a price-driven recovery is exactly the kind the network is engineered to erode.

The thermostat re-engages: difficulty’s 2026 round trip

Difficulty is Bitcoin’s thermostat. Every 2,016 blocks (about two weeks) the protocol compares how fast those blocks actually arrived against the 10-minute target and resets the mining difficulty to drag the average back toward ten minutes. Satoshi built it so that no matter how much or how little computing power shows up, blocks keep landing at a steady pace. It is automatic, hard to game, and completely indifferent to anyone’s margins.

The adjustment is bounded and formula-driven, not discretionary. Each retarget can move difficulty by at most a factor of four in either direction, a clamp that has never been hit in practice, and it keys off the timestamps written into block headers rather than any direct reading of hashrate, which is why every hashrate figure you see is an estimate inferred from how fast blocks arrive. Early in a two-week window those estimates are noisy; they sharpen as blocks accumulate, which is why trackers often disagree on the size, and occasionally the sign, of the next move until the epoch is nearly complete.

In 2026 that thermostat did something it had almost never done: it fell, hard, by choice. As the price squeeze and the pull of AI computing drove operators to switch machines off or repoint their power, difficulty dropped to a low of 124.93T on June 14, a 10.09 percent cut and the largest of the year. For only the second time in history, it then spent months running below its year-earlier level, a shift CoinDesk flagged as the network shrank 14 percent off its 2026 high. Then the price rally reversed the flow. Machines came back on. Difficulty climbed: up 0.99 percent on August 8, up 1.31 percent on September 6 (the eighth increase of the year, per news.bitcoin.com), then up 4.16 percent on September 19, the biggest single jump since June. On October 3 it printed essentially flat, settling at 132.72T.

Read down that last column and you can watch the network claw back the summer’s decline. From the June floor, difficulty is up roughly 6 percent, and it now sits only about 15 percent below the November 2025 record of 155.97T. Every one of those upward steps raised the denominator of the hashprice formula, which means every one of them quietly took back a slice of the price-driven recovery before miners ever saw it in their wallets.

Why the network, not the Fed, is the ceiling

Here is the mechanism in plain terms. When Bitcoin’s price rises, hashprice rises with it, and mining becomes more profitable at the margin. Profitable mining pulls idle or newly built machines online. More machines mean more hashrate, and about two weeks later the thermostat responds by raising difficulty. Higher difficulty lowers hashprice. The system chases its own tail back toward the point where the least efficient machine still running roughly breaks even.

That tail-chasing is what makes Bitcoin mining a commodity business in the textbook sense, with one twist. Unlike almost any other commodity, the supply of the product (new bitcoin) is fixed by the protocol no matter how many producers show up. Extra competition cannot add supply; it can only raise the cost of winning the same fixed reward. Difficulty therefore behaves like an endogenous marginal cost, a cost of production that rises and falls with revenue, forever dragging the whole fleet back toward the breakeven of the least efficient machine still worth running. A miner does not get to keep an above-market margin for long; the network notices, and competes it away.

JPMorgan’s Nikolaos Panigirtzoglou described the downside of that loop precisely: in a client note he wrote that “when bitcoin trades below its production cost, higher-cost miners power down, the hashrate declines, and difficulty adjusts lower.” In 2026 that sentence has been running in reverse. Bitcoin climbed back above the roughly $78,000 all-in production cost his team estimates, higher-cost miners powered back up, hashrate rose, and difficulty adjusted higher. It is the same physics read forward or backward; the summer was the decline, the autumn is the recovery, and the thermostat governs both. His team pegs the sensitivity of difficulty to price (a beta of about 0.62), which is another way of saying the network recaptures most of a rally over time.

This is why framing hashprice as a bet on the Federal Reserve, popular as it is, misses the binding constraint. Rates matter, of course: lower rates tend to lift risk assets, Bitcoin among them, and a higher Bitcoin price lifts hashprice. But that chain is indirect, and it runs through a variable (the Bitcoin price) that no miner controls and the Fed does not target. Difficulty is the direct lever, and it responds specifically to mining activity, competing away exactly the margin a rally creates. The Fed can nudge the numerator by accident; the network moves the denominator on purpose.

Fred Thiel, chief executive of MARA, put the end state bluntly. Mining, he told CoinGeek, “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.” The floor is energy cost; the ceiling is difficulty. A miner’s whole job is to live in the gap between them, and the gap is what the network is forever trying to close. It is a very different wager from the one a proof-of-stake validator makes, where the reward schedule is set by code rather than by a two-week hardware race; the contrast between a validator’s and a miner’s block-reward economics is really a contrast in who controls the denominator.

BTC hashprice versus USD hashprice

The cleanest way to watch the thermostat at work is to quote hashprice in Bitcoin instead of dollars. In dollar terms, hashprice is up roughly 45 percent from the June low. In Bitcoin terms it has barely moved, hovering near 0.00047 BTC per PH/day, close to where it sat at the bottom.

That gap is the whole story in one data point. The recovery is a fiat-price event, not a network event. Strip out the rising Bitcoin price and a petahash of computing power is earning about the same number of satoshis per day as it did at the lows, because the thing that actually sets sats-per-hash (the reward split across total hashrate) is governed by difficulty, and difficulty has climbed right alongside the price. BTC-denominated hashprice is, in effect, mining’s real yield: what the work earns in the asset itself, before the currency translation flatters the number.

For operators, the distinction is practical, not academic. A miner who believes in Bitcoin over the long run and holds the coins cares about the Bitcoin figure, and that figure says the autumn changed far less than the dollar headlines suggest. A miner who sells production to pay dollar-denominated power bills cares about the dollar figure, and for that operator the rally is real income, for exactly as long as the price holds and difficulty does not fully catch up.

From hashprice to break-even: the energy-adjusted view

Hashprice is revenue, not profit. To turn it into a margin you need one more input, the efficiency of your machines, measured in joules per terahash (J/TH), and the price you pay for electricity. The relationship is simple enough to run on a napkin: a machine breaks even when its revenue equals its power cost, which works out to a break-even power price (in dollars per kWh) equal to hashprice divided by (24 times J/TH).

At $40 per PH/day, that formula produces the following break-even power prices by machine. Below the listed price the machine earns a gross margin; above it, the machine loses money on electricity alone, before you even count hosting and depreciation.

Put concrete numbers on it. An operator running newer S21 XP machines (13.5 J/TH) at 5 cents a kilowatt-hour collects about $40 of revenue per petahash-day against roughly $16 of power, a gross margin near 60 percent. Swap in five-year-old S19j Pro units (29.5 J/TH) on that same 5-cent power and the power bill jumps to about $35, leaving a margin so thin that a single difficulty uptick can erase it. The same contrast explains why the hosting-versus-self-mining choice matters so much: a hosted miner pays an all-in rate (often 6 to 8 cents) that bundles the host’s margin, while a self-miner who owns cheap generation can sit far down the cost curve. Hashprice is identical for both; what they keep is not.

This is where the single global hashprice number splinters into a thousand local outcomes. The same $40 that leaves a newest-generation hydro machine comfortably profitable at a Gulf or West Texas power price under 6 cents a kilowatt-hour will sink a five-year-old S19j Pro anywhere electricity costs more than about 5.6 cents. On high-cost industrial grids (much of Western Europe sits above 20 cents a kilowatt-hour), nearly the entire fleet is underwater at $40, which is precisely why the network’s hashrate keeps migrating toward the cheapest power on earth. Hashprice tells you the revenue; geography and chip age decide whether that revenue is a profit.

The security budget: what the ceiling means for Bitcoin itself

Zoom out from the single miner and hashprice takes on a second meaning: it is the price of Bitcoin’s security. Multiply hashprice by total hashrate (or, equivalently, multiply the daily block reward by the Bitcoin price and annualize) and you get the total revenue flowing to miners, which is roughly $14 billion a year at current levels. That figure is the economic cost of attacking the network, because a would-be 51 percent attacker has to out-muscle the honest miners collectively earning it. When hashprice recovers, the security budget grows; when the thermostat competes it back, the budget stops growing.

There is a structural worry buried in that budget. Almost all of it, well over 99 percent, is the block subsidy; transaction fees have run under 1 percent of the reward for most of 2026. The subsidy halves roughly every four years, with the next halving due in 2028 to cut it from 3.125 to 1.5625 BTC. At a constant Bitcoin price, that single event halves the security budget. The long-run bet is that transaction fees grow to fill the gap, but 2026’s fee data shows how far there is to go. Taproot and the turn toward Bitcoin-native self-custody were supposed to seed more on-chain activity; the flat fee line suggests that dividend is still mostly ahead.

This is the part of the hashprice story that reaches beyond mining profit-and-loss. Every holder, every exchange, every spot ETF has a stake in whether the number stays high enough, through price now and eventually through fees, to keep paying for the hashrate that makes Bitcoin expensive to attack. A permanently low hashprice is not just a miner problem; it is a security problem for everyone who owns the asset.

Hedging the ceiling: hashrate forwards and hashprice derivatives

If difficulty is going to keep clawing back price gains, miners need a way to lock in revenue they can see today. That demand built an entire derivatives market on top of hashprice. Luxor runs over-the-counter hashprice forwards out to twelve months, and Bitnomial lists CFTC-regulated hashrate futures (ticker HUP) in one-petahash monthly contracts. A miner who sells a forward at today’s hashprice locks that revenue in regardless of what difficulty does next; the counterparty takes the opposite view.

The shape of that forward curve is itself a forecast, and in late 2026 it tells a consistent story. The six-month forward has been trading around $37 to $38 per PH/day, below the roughly $40 spot. A forward below spot (backwardation) means the market, in dollars and cents, is pricing the recovery to fade, which is precisely the outcome the difficulty feedback loop produces. The implied cost of capital embedded in those contracts has run in the 6 to 13 percent annualized range, a real-world read on how expensive it is to de-risk a mining operation against its own network.

One regulatory nuance matters here, and it is a useful tell about how US authorities see mining. The act of mining is not a security; the derivatives written on its output are commodity contracts, overseen by the Commodity Futures Trading Commission rather than the SEC. That is the same jurisdictional line now shaping where the rest of crypto’s derivatives business is allowed to operate.

The escape hatch that competes for the same megawatts: AI and HPC

There is a second force bending the difficulty curve in 2026, and it is the reason the summer decline happened by choice rather than by distress. The same megawatts, power contracts, substations, and cooling that run Bitcoin ASICs can run AI accelerators, and for now the economics of doing so are lopsided. Needham analyst John Todaro summed it up to CoinGeek: “the revenue per megawatt and EBITDA margins are far higher for HPC and AI colocation than for mining.”

The capital has followed that math. CoinShares estimates more than $70 billion in cumulative AI and high-performance-computing contracts announced across the public mining sector, and Butterfill has projected that some listed miners could draw up to 70 percent of their revenue from AI by the end of 2026, up from roughly 30 percent, a shift he detailed to news.bitcoin.com. When an operator repoints a data hall from hashing to GPUs, that hashrate leaves the network, and two weeks later difficulty falls and hashprice rises for whoever stayed. AI is thus both a drain on hashrate and, paradoxically, a support for the hashprice of the miners who remain. The demand pulling those megawatts toward GPUs is the same compute boom that has minted and re-minted AI compute tokens all year.

It also reframes the ceiling. For a miner with a credible AI pivot, the relevant comparison is no longer just hashprice against a power bill, but hashprice against what that same megawatt could earn leased to an AI tenant. On that scoreboard a $40 hashprice often loses outright. The difficulty thermostat caps mining revenue; the AI bid sets the opportunity cost of staying in mining at all.

Hashprice, taxes, and the SEC

For US miners, the regulatory picture around hashprice is unusually settled. In March 2025 the SEC’s Division of Corporation Finance stated that proof-of-work mining, whether solo or through a pool, is not the offer or sale of a security, because miners rely on their own computing effort rather than the managerial efforts of others. That lifted a long-standing cloud: earning hashprice does not, by itself, pull a miner into securities law.

The derivatives are a different regulator’s problem, as noted above, and mined Bitcoin is still property for tax purposes, taxed as ordinary income at its fair value when it is received and again for any gain when it is sold. None of that changes the core number; it only determines how much of it a miner keeps.

The broader policy backdrop is less tidy. A thinly staffed SEC, and the unresolved question of who referees crypto markets at all, left much of 2026’s rulemaking in limbo. The agency that cleared mining of securities concerns is itself short-handed, a gap that has shaped crypto policy all year. For hashprice specifically, though, the regulatory risk is low; the economic risk, set by price and difficulty, is where all the action is.

What moves hashprice next

Three forces will set hashprice through the rest of 2026, and only one of them works in miners’ favor. The first is the Bitcoin price, the numerator. To show the ceiling one more time: at today’s roughly 950 EH/s, a full retest of the $126,080 record would lift hashprice only to about $60 per PH/day, and a climb back to $100,000 to about $47. Those are not small numbers next to a June low near $27.70, but they are what the formula pays before a single extra machine switches on.

The second force is the one that always answers back. The next difficulty retarget, estimated around October 17, points to roughly +0.62 percent (CoinWarz), and any sustained price strength would pull idled and newly built rigs online, sending later retargets higher still. CoinShares has framed the path as conditional: hashrate only resumes its climb toward 1.8 to 2 zettahashes if Bitcoin recovers durably, and every unit of hashrate that returns is a unit of hashprice competed away. The forward market already reflects the tension, pricing six-month hashprice below spot.

The third force is the calendar. The Federal Reserve’s next decision lands on October 27 and 28, the kind of macro event that moves the price leg the recovery rests on, and beyond it sits the 2028 halving, which will cut the subsidy, and with it the single largest input to hashprice, in half. For now, at around $40 per PH/day, hashprice sits in a fragile middle: high enough to keep most efficient fleets alive, low enough that the next move depends entirely on whether price can keep outrunning a network engineered to catch up.

Frequently Asked Questions

What is hashprice in simple terms?

Hashprice is the revenue a Bitcoin miner can expect to earn per unit of computing power per day, quoted in US dollars per petahash per second per day ($/PH/day). It bundles the block subsidy, transaction fees, the Bitcoin price, and network difficulty into one number, so a miner can compare expected revenue directly against electricity costs.

Why did hashprice recover in 2026 while Bitcoin-denominated hashprice did not?

The recovery was driven almost entirely by Bitcoin’s price rising from the low $60,000s to the mid $80,000s, and the price sits in hashprice’s numerator. Measured in Bitcoin, hashprice barely moved, near 0.00047 BTC per PH/day, because network difficulty climbed alongside the price and kept the satoshis earned per unit of hashrate roughly flat.

How does mining difficulty cap hashprice?

Difficulty, or equivalently hashrate, sits in the denominator of the hashprice formula. When a rising price makes mining profitable, more machines come online and raise hashrate; about two weeks later the network’s automatic difficulty adjustment increases, which lowers hashprice. This feedback loop competes away the margin a price rally creates, so difficulty acts as a ceiling on how much of any rally reaches miners.

What hashprice do miners need to be profitable?

It depends on machine efficiency and power cost. At $40 per PH/day, a top-tier machine around 13.5 J/TH breaks even near $0.12 per kWh, while an older 29.5 J/TH rig breaks even near $0.056 per kWh. Operators with power under 5 to 6 cents per kWh are profitable across most of their fleet, while those on expensive grids can lose money even on efficient hardware.

Where can I track hashprice and difficulty myself?

Hashrate Index publishes the most-watched hashprice index and weekly roundups; CoinWarz and mempool.space track network difficulty, hashrate, and the estimated next retarget; and CoinGecko tracks the Bitcoin price that drives the numerator. Watching all three together shows whether a move in hashprice is a price event or a network event.

By Yuki Tanaka, mining and markets desk, HOGE Wire.