Bitcoin mining difficulty set to fall about 4.8% on 13 November, after 375 days below its record

Bitcoin blockchain and earns a reward, which makes rewriting history very expensive.">mining difficulty is on course to fall by about 4.8% at the next adjustment, due at block 975,744 on Friday 13 November around 12:25 UTC. That was the estimate published by the open-source explorer mempool.space at 11:53 UTC on Sunday 8 November: -4.79%, with 688 blocks still to be mined in the current two-week window. A cut of that size would be the biggest since 11 July and the first fall after two rises in a row.

The figure is not final. It moves with every block, and it stood near -2.8% on 2 November. What does not move is the backdrop: the network’s difficulty has now spent 375 days below its record of 29 October 2025, the longest such stretch in bitcoin’s history, while miners earn about $33 to $35 a day per unit of computing power, not far above the record low of June.

Bitcoin traded near $75,910 at 11:41 UTC on Sunday, down 2.9% over seven days, according to our market snapshot.

How the bitcoin difficulty adjustment works

Every 2,016 blocks, roughly every two weeks, the bitcoin protocol compares how long those blocks actually took with the target of ten minutes each. If they came too fast, mining gets harder. If they came too slowly, it gets easier. The aim is a steady issuance rate whatever the amount of computing power, or hashrate, pointed at the network. Our guide to hashrate and the difficulty adjustment explains how both numbers are estimated.

Difficulty itself is just a number. On 29 October it was set at 148.36 trillion (148.36T), up 5.54%. Two weeks earlier, on 16 October, it had risen 5.92% to 140.58T, a move we covered in our report on the October difficulty jump.

Bitcoin mining difficulty estimate: the numbers behind a 4.8% cut

The estimate is simple arithmetic. Since the retarget of 29 October, 1,328 blocks have been mined. At ten minutes per block, about 1,396 should have been mined by 11:53 UTC on Sunday. The network is running about 4.9% behind schedule, and blocks in this window have averaged 10 minutes 31 seconds instead of ten minutes.

If the remaining 688 blocks keep that pace, they will take about five days, which puts the retarget on Friday 13 November around midday UTC. The protocol would then lower difficulty by enough to bring the next window back to ten minutes.

Applied to the current 148.36T, a 4.79% cut would bring difficulty to about 141.3T by our calculation. That is still well above the year’s low of 124.93T, set on 14 June after a 10.09% fall according to Luxor’s Hashrate Index lookback for June, and about 9.4% below the record.

The estimate has grown during the week. On 2 November, the Hashrate Index newsletter put it at -2.8% and warned that early predictions are “shaky because of short-term variance”. On Saturday 7 November, French outlet Cointribune reported about -3.7% from mempool.space. By Sunday midday it stood at -4.79%. A run of fast blocks in the next five days could shrink it again; a slow run could take it past -5%.

Source: mempool.space list of difficulty adjustments, dates in UTC from block timestamps. Of the 22 adjustments made so far in 2026, 11 were cuts by our count. The four largest were -11.16% in February, -10.09% in June, -7.76% on 20 March and -5.00% in July. A cut of about 4.8% would rank fifth.

375 days below the record, longer than 2011-2012

The record still stands at 155.97T, set at block 921,312 on 29 October 2025, according to mempool.space block data. Counting calendar days, Sunday 8 November is the 375th day without a higher reading.

The previous longest gap ran from 1 August 2011 to 30 July 2012: 364 days, during the period when bitcoin’s price lost more than 90% after its first bubble. By our count on the blockchain.info daily difficulty series, that mark was passed around 28 October 2026. The +5.54% rise of the following day brought difficulty within 4.9% of the record but not above it.

Source: our calculation on the blockchain.info daily difficulty series, 8 November 2026; the current record date uses mempool.space block data. Cointribune and Cryptoast reach the same 364 days for 2011-2012 and count 253 days after the China ban, one more than our daily series gives.

Bitcoin author Saifedean Ammous argued, in remarks reported by Cryptoast on 8 November, that neither the 95% price crash of 2011 nor the Chinese ban had been as severe as the current stretch. The comparison holds on duration. It is less obvious on depth: in 2011 the network was tiny and the price collapse was far deeper than bitcoin’s fall of about 12% this year, measured on 2 November.

Part of the explanation lies in where the record was set. October 2025 was a peak for hashrate, reached before the price slid. Then came a run of cuts: -11.16% in February, -7.76% in March, -10.09% in June. Each time hashrate climbed back, difficulty fell short of the October 2025 peak; the highest reading of 2026 is the 148.36T set on 29 October.

After a cut of about 4.8% on 13 November, the new difficulty would sit about 9.4% below the record. The following retarget comes 2,016 blocks later, around 27 November if blocks arrive every ten minutes. Beating 155.97T then would need a sustained hashrate well above the late-October highs, at a time when miners are being paid less. The 375 days look set to keep growing into December at least.

Why blocks slowed: hashrate fell back to about 1 ZH/s

Difficulty reacts to hashrate, and hashrate reacts to money. In October, mempool.space daily estimates of the network’s computing power passed 1.1 zettahashes per second (ZH/s, a billion trillion hashes per second) on several days, with a high of 1,228 exahashes per second (EH/s) on 24 October. Difficulty followed with two rises of more than 5%. That level did not hold.

By 2 November, the seven-day average hashrate had slipped to 1,062 EH/s from 1,091 a week earlier, according to the Hashrate Index roundup of 2 November. On Sunday at 11:53 UTC, mempool.space put current hashrate at about 998 EH/s. Daily figures are noisy, because some days are luckier than others, but the direction since late October is down.

The network is not shrinking in any dramatic sense. Hashrate is still around 1 ZH/s, the weekly figure having returned to that level after eight months below it, Cointribune reported on 7 November, citing Hashrate Index data. What changed is that machines switched on in October at higher prices have been switched off again, or pointed at other work.

The price explains much of it. Bitcoin stood at $76,752 when Hashrate Index published its roundup at 19:53 UTC on 2 November, about 12% down for the year, and near $75,910 at 11:41 UTC on Sunday. Machines that paid their way in mid-October at higher prices stop doing so a few thousand dollars lower.

Mining remains concentrated. Over the seven days to Sunday 11:53 UTC, Foundry USA found 225 of 933 blocks (24.1%), AntPool 176 (18.9%) and F2Pool 170 (18.2%), per mempool.space. The top three pools took 61.2% of blocks. The 933 blocks themselves are about 7% short of the 1,008 a full week should produce.

The miners’ squeeze: hashprice near $33, fees under 1%

Miners measure their income in hashprice: the dollars earned per petahash per second (PH/s) of computing power per day. Hashrate Index put it at $32.86 on 2 November, down 8.7% in a week and below its 30-day average of $37.49. Our own calculation from mempool.space data on Sunday, using its bitcoin price of $75,896 at 11:45 UTC, gives about $34.4. Either way, the range is about $33 to $35.

That is not far above the record low of $27.74, reached on 6 June 2026, per Luxor. June’s average of $30.37 was the lowest monthly figure on record.

Transaction fees barely move the needle. The last 144 blocks up to block 975,056 on Sunday, about one day of mining, paid 452.86 BTC in total, of which 2.86 BTC were fees: 0.63%. Hashrate Index counted 0.55% for the week to 2 November. Almost all of a miner’s revenue is the fixed block subsidy of 3.125 BTC, which halved in 2024 and will halve again around 2028, as our explainer on the bitcoin halving details.

Hashrate Index also translates hashprice into an energy price a miner can afford. Fleets of the newest machines, using less than 14 joules per terahash, earn the equivalent of $116 per megawatt-hour. Older machines in the 25-38 J/TH band earn $43/MWh. A miner running old hardware on power that costs more than that loses money by mining, which is why those machines are the first to stop.

Riot sold 20,264 BTC in nine months

Riot Platforms, one of the largest listed US miners, gave a clear view of the sums in its third-quarter results filed with the SEC on 4 November. In the first nine months of 2026 the company sold 20,264 bitcoin for about $1.5 billion, “primarily to fund data center development”. It held 2,418 BTC at 30 September.

Riot mined 1,637 BTC in the third quarter, against 1,406 a year earlier. Its average cost to mine one bitcoin in the quarter was $39,265 excluding the depreciation of its machines. Including that depreciation, the figure was $78,805, above the $70,855 average value of each coin it mined in the quarter and above the $75,910 bitcoin traded at 11:41 UTC on Sunday. The gap shows why a miner can keep its fleet running while the business of mining, counted in full, does not pay for new machines.

The money is going to data centres. Riot signed a 15-year lease at its Corsicana site in Texas for 910 megawatts, worth about $28 billion, and says it now has 1,151 MW leased. “We do not currently anticipate issuing equity to fund our announced projects,” chief executive Jason Les said in the release.

Riot is not alone. IREN and Core Scientific have set year-end dates to leave bitcoin mining, and TeraWulf now earns a small share of its revenue from it, as described in our report on miners moving to AI.

Tracking this shift month by month has become harder. Riot moved from monthly to quarterly production reports at the start of 2026, and CleanSpark’s update for September, when it mined 529 BTC, was its last monthly one. There are no October production figures from either. For now, the most reliable live gauge of miners’ behaviour is the network itself: block times, hashrate and the difficulty estimate.

What the 13 November cut changes, and what to watch

A lower difficulty means each machine finds a slightly larger share of blocks. If hashrate holds where it is, a 4.79% cut raises the number of bitcoin earned per unit of computing power by about 5%. At a stable bitcoin price, hashprice would rise by about the same amount after Friday. That is our reading of the mechanics, not a forecast.

It cuts both ways. Cheaper blocks can draw idle machines back, as happened after the June cut, which was followed by a 7.15% rise on 27 June. If enough return, blocks speed up and the following retarget, around late November, goes up again. That is the pattern of 2026 so far: half the adjustments were cuts, half were rises, around a hashrate that has gone sideways near 1 ZH/s.

For someone holding bitcoin, the adjustment changes nothing directly. The supply schedule stays the same, and a network running at about 1 ZH/s still has close to the most computing power it has ever had. For miners and their shareholders, it is a small relief on revenue in a year when Hashrate Index counted 17 of 18 mining stocks down in the week to 2 November.

Things to watch before Friday:

• The mempool.space estimate, which can move by a percentage point in a day as block luck changes.
• The seven-day average hashrate: a drop below 1 ZH/s would confirm that more machines are switching off.
• The bitcoin price around the US inflation figures (CPI) due on Tuesday 10 November at 13:30 UTC.
• Any new sales of bitcoin by listed miners in their filings.

The retarget itself should land at block 975,744 around midday UTC on Friday 13 November, give or take a few hours. For how mining works and what a miner’s costs look like, see our guide to bitcoin mining in 2026.