Bitcoin could lower mining difficulty by roughly 16% when its next adjustment arrives around July 26, handing the machines that remain online a considerably larger expected share of the network’s rewards while doing almost nothing to resolve the expensive power contracts, debt obligations and strategic pressures pushing some of the industry’s largest companies away from mining.
The network already lowered difficulty by 5% at block 957,600 on July 11, bringing it to 127.17 trillion. Hashrate Index reported that hashprice, the daily revenue miners expect from one petahash per second of computing power, stood at $30.88 per PH/s/day on July 13, with a seven-day average of $30.39.
That level was at or below breakeven for many operators depending on their power costs and machine models. While hashprice had recovered from the $27.60 level recorded around the beginning of July, it remains 37% below its October 2025 peak near $49.40.
Bitcoin’s protocol can automatically lower the computational burden required to produce blocks, but it can’t renegotiate a miner’s electricity contract, refinance convertible notes, restore the value of aging machines, or make volatile mining revenue more attractive than a multibillion-dollar AI lease.
Bitcoin can repair its schedule, not miner balance sheets
Bitcoin adjusts difficulty every 2,016 blocks according to how long the previous 2,016 took to produce, lowering the requirement when hashrate leaves and blocks slow down, then raising it again when machines return and blocks begin arriving too quickly.
The mechanism is designed to pull the average block interval back toward ten minutes, and when difficulty falls, each surviving miner receives a larger expected share of the same 3.125 BTC subsidy for every unit of hashrate it contributes.
In that sense, the mechanism works just as intended, but the value of its relief still depends on three other variables that difficulty cannot control: Bitcoin’s market price, transaction fees, and the speed with which competing machines return after conditions improve.
The situation since June shows why the benefit can disappear almost as quickly as it arrives, because difficulty fell 10.09% in mid-June, rose 7.15% on June 26 as hashrate returned, then fell another 5% on July 11 after computing capacity weakened again.
Across 2026, eight of Bitcoin’s first 14 adjustments were negative, while compounded difficulty declined approximately 14.22% from the January 8 peak of 146.47 trillion. These numbers show a network that has shed capacity faster than it has sustainably replaced it.
Transaction fees have provided almost no cushion, with miners collecting approximately 2,914 BTC in total rewards during the week to July 13 while fees represented only 0.69% of block rewards.
A downward retarget can therefore improve revenue when Bitcoin’s price and transaction fees remain stable. However, it can be neutralized by another decline in BTC, erased when efficient competitors switch back on, or rendered irrelevant for companies whose financing costs and electricity expenses already exceed what a roughly $31 hash price can support.
A double-digit cut would be relief, not a rescue
The estimate for the current epoch initially pointed in the opposite direction, with Luxor projecting a 2.74% increase on July 13 when blocks over the preceding 24 hours were arriving in an average of nine minutes and 44 seconds.
Block production subsequently slowed enough for the estimate to swing toward a substantial reduction, and although the final percentage will remain volatile until the closing blocks of the epoch, the reversal is important because it indicates that capacity has continued leaving faster than it can return.
A double-digit reduction would immediately improve the economics of every machine that remains online, but most of the benefit would flow toward operators with the newest equipment, the cheapest electricity, the strongest access to capital and enough balance-sheet flexibility to restart fleets without selling assets under pressure.
That creates a less intuitive possibility in which aggregate hashrate recovers while the industry underneath it becomes more concentrated, as efficient fleets switch back on, low-cost operators gain network share, aging machines remain permanently dark and a small number of large additions conceal many smaller exits.
In other words, a healthier block interval doesn’t necessarily indicate a healthier mining sector, an issue CryptoSlate previously examined in its analysis of how public miners’ BTC sales and AI pivots could reshape Bitcoin’s security base.
CleanSpark is the best example of an operator that should be capable of capturing difficulty relief, because its June operating update reported 50 EH/s of operational hashrate, an average operating hashrate of 42.6 EH/s, and peak deployed-fleet efficiency of 16.07 joules per terahash.
The company nevertheless produced only 614 BTC, down from 671 BTC in May, while selling 179 BTC at spot and another 250 through call exercises at an average realized price of $69,056.
CleanSpark’s reported holdings still rose to 13,924 BTC, but purchases and a delta-neutral basis trade more than offset the 429 coins that left during the month, while 1,719 BTC were posted as collateral or recorded as receivables connected to derivative transactions.
That distinction between reported holdings and immediately deployable liquidity was explored in CryptoSlate’s previous coverage of the liquidity test inside miner Bitcoin reserves, because a large treasury doesn’t necessarily provide the same financial flexibility when part of it is restricted, pledged, receivable, or already serving another balance-sheet function.
MARA is the more dramatic counterpoint, having sold 20,880 BTC for approximately $1.5 billion during the first quarter of 2026 while reporting a $1.26 billion net loss on $175 million of revenue, cutting roughly 15% of its workforce and recognizing $45.9 million in restructuring charges.
According to the company’s first-quarter results, MARA used its Bitcoin treasury as a source of liquidity while reducing its large-scale ASIC purchases, and a separate company disclosure detailed the sale of 15,133 BTC connected to a $1 billion repurchase of convertible senior notes.
A difficulty reduction can improve the revenue of MARA’s remaining machines, but it’s unlikely to reverse a strategy in which mining equipment, Bitcoin reserves, debt management and convertible AI infrastructure are being evaluated together.
| Company | Mining position | Bitcoin and balance-sheet signal | AI or data-center position | What lower difficulty means |
|---|---|---|---|---|
| CleanSpark | 50 EH/s operational, 42.6 EH/s average operating hashrate and 16.07 J/TH peak fleet efficiency in June. | Produced 614 BTC and sold 429 BTC through spot transactions and call exercises; holdings reached 13,924 BTC after additional treasury transactions. | Developing and commercializing AI and high-performance-computing campuses, although no major contracted AI revenue was disclosed in the June mining update. | One of the operators best positioned to capture immediate relief, although continuing BTC monetization shows that efficiency has not removed liquidity pressure. |
| MARA | Mining remains a large operating business, but management has reduced its emphasis on major new ASIC purchases and is restructuring parts of the operation. | Sold 20,880 BTC for about $1.5 billion during the first quarter while reporting a $1.26 billion net loss. | Positioned most non-hosted capacity so it can support AI and critical IT workloads as part of a broader digital-infrastructure strategy. | Improves unit mining economics but is unlikely to change the company’s debt reduction, restructuring, or infrastructure-conversion strategy. |
| Riot Platforms | 42.5 EH/s deployed and 36.4 EH/s average operating hashrate during the first quarter. | Produced 1,473 BTC and sold 3,778 BTC for net proceeds of $289.5 million at an average price of $76,626. | An AMD data-center lease established contracted critical IT capacity at its Rockdale site, with expansion rights extending the potential deployment. | Riot can capture relief through its scale, but BTC sales and contracted compute revenue reduce the importance of difficulty to the broader investment case. |
| TeraWulf | Bitcoin mining remains part of the company’s infrastructure portfolio, although its strategic growth narrative is increasingly tied to contracted AI capacity. | The company’s financial future is becoming less dependent on self-mined Bitcoin and short-term changes in hashprice. | Signed a 20-year Anthropic lease covering approximately 401 MW of critical IT load and roughly $19 billion of expected contracted revenue. | Lower difficulty still benefits mining operations, but it has become secondary to the economics of a multidecade AI contract. |
| Hut 8 | Operates a power-first infrastructure platform in which Bitcoin mining is one use of energy capacity rather than the sole business. | Bitcoin exposure remains part of the company, but the value of its power, land and interconnections is increasingly measured through long-duration leases. | A second 352 MW Beacon Point lease increased contracted capacity at the campus to 704 MW and campus-level base-term contract value to $19.6 billion. | Difficulty relief has become close to irrelevant to the largest part of the company’s emerging contracted-revenue story. |
Data note: Mining figures, treasury activity and contract values are tied to the reporting periods and announcement dates in the linked primary sources; they should not be interpreted as real-time operating metrics.
The uneven benefit from a difficulty cut ultimately comes down to electricity, because an efficient machine connected to low-cost power can return almost immediately when hashprice improves, while an older fleet tied to an expensive contract can remain unprofitable even after a double-digit reduction.
A 2026 working paper examining the Texas electricity market found that mining load declines as electricity-sector costs rise, but that the strength of the response depends on hashprice, with higher expected mining revenue allowing machines to remain online at higher wholesale power prices.
The researchers also warned that treating mining as a permanently available demand-response resource may overstate how much flexibility miners can actually provide, because their willingness to curtail depends on the relationship between electricity prices and the revenue available from hashing at that moment.
Applied to the coming adjustment, lower difficulty will allow some machines to tolerate moderately higher power prices, but it will not rescue a fleet whose contracted electricity cost already exceeds what the combination of BTC price, fees, and network competition can support.
The pressure is especially visible in Texas, where summer peaks can encourage operators to shut down during expensive hours while also managing exposure to future transmission charges, an infrastructure conflict CryptoSlate examined in its report on why miners have until 2027 to prove their value to an overloaded US power system.
AI has changed what a mining site is worth
The most important competition facing Bitcoin miners may no longer come from another mining company, because a site with available electricity, suitable land, cooling infrastructure, and a secured grid interconnection has become a scarce asset that can support several forms of high-density computing.
Bitcoin mining can be deployed relatively quickly, but its revenue changes continuously with Bitcoin’s price, transaction fees, difficulty, and the efficiency of competing machines. A hyperscaler or AI-infrastructure lease may take longer and require considerably more capital, but provide predictable payments extending for a decade or more.
CryptoSlate previously reported that miners had announced more than $70 billion in AI and high-performance-computing agreements, with some operators expected to derive as much as 70% of their revenue from those workloads by the end of 2026. Most of the expansion was being financed through a combination of debt and Bitcoin sales.
The transition has only accelerated, with TeraWulf’s Anthropic agreement and Hut 8’s expanded Beacon Point lease showing that the strongest alternative to volatile mining revenue is no longer theoretical, but represented by signed contracts carrying tens of billions of dollars in base-term value.
As CryptoSlate noted in its analysis of why miners’ real prize may be their access to power, AI demand changes the opportunity cost of every megawatt assigned to Bitcoin. This is because keeping a mining fleet online now means comparing its expected revenue not only with the electricity bill, but with the potential value of converting the same site into contracted compute infrastructure.
A difficulty reduction slightly improves that calculation in Bitcoin’s favor, but it doesn’t eliminate the attraction of a 15-year or 20-year agreement supported by an investment-grade tenant, particularly when mining revenue remains near the sector’s production-cost floor.
What the next adjustment will reveal
Three outcomes will matter once the retarget lands, because the initial percentage change won’t say anything about the industry, but about the behavior of hashrate in the days that follow it.
A smaller reduction than the current estimate would suggest that hashrate returned during the closing stages of the epoch and that competition is already rebuilding, limiting the duration of the margin improvement available to miners that remained online.
A large reduction followed by a rapid recovery in hashrate would indicate that efficient operators had machines waiting for better economics, allowing the strongest fleets to restart while weaker operators remained distressed and setting the network up for another upward adjustment during the following epoch.
A large reduction followed by persistently weak hashrate would point toward more permanent fleet retirements, continued consolidation and an accelerating migration of power infrastructure toward workloads outside Bitcoin.
The most useful signals will therefore be hashprice after the adjustment, production updates from public miners, miner-to-exchange transfers, changes in mining-pool share and the estimated direction of the following epoch, rather than the block-time correction alone.
Bitcoin’s difficulty mechanism will continue doing what it was designed to do, restoring block production toward its ten-minute target and increasing the expected rewards available to the machines that remain online.
But it can’t determine whether those machines belong to a broad and financially healthy mining industry or to a shrinking group of efficient operators consolidating network share while competitors sell Bitcoin, retire equipment, and redirect their electricity toward AI.
