Bitcoin mining difficulty fell to 126.23 trillion as of August 1, 2026, down 19.1% from its all-time high of 155.97 trillion set in November 2025. The metric has dropped below its year-ago level for only the second time in Bitcoin's 17-year history — the first being China's mining ban in mid-2021...
"You get a lot more money per electron if you're doing it for AI than for Bitcoin mining." — Fred Thiel, CEO, MARA Holdings
Bitcoin mining difficulty fell to 126.23 trillion as of August 1, 2026, down 19.1% from its all-time high of 155.97 trillion set in November 2025. The metric has dropped below its year-ago level for only the second time in Bitcoin's 17-year history — the first being China's mining ban in mid-2021. Unlike 2021, this decline was not triggered by a single regulatory shock. It is the product of compounding economic pressure: production costs exceeding spot price, a structural capital migration toward AI data center contracts, and regional power disruptions across Texas, Iran, and other mining hubs.
Publicly listed miners sold more than 32,000 BTC in Q1 2026 alone — exceeding their combined sales for all of 2025 and surpassing the 20,000 BTC liquidated during the 2022 Terra-Luna collapse. Hashprice, the standard measure of daily miner revenue per unit of computing power, touched $27.66 per petahash per day in late June, a five-year low. Luxor's forward market prices an average of $31.85 through December, signaling that participants expect little recovery this year. Meanwhile, publicly listed miners have committed to over $70 billion in cumulative AI and high-performance computing contracts, with projections showing AI-related services could constitute up to 70% of these companies' revenue by year-end.
Bitcoin mining difficulty has recorded three consecutive negative adjustments in 2026: a 11.16% drop to 125.86 trillion on February 7, a 10.09% decline from 138.96 trillion to 124.93 trillion in June, and a further 5% reduction in July. The current reading of 126.23 trillion sits 1.1% below the 127.62 trillion recorded one year earlier, according to CoinWarz data.
Network hashrate has fallen below 1 zettahash per second (ZH/s) and remains there, a threshold the network first crossed in late 2025. As of Q2 2026, implied hashrate stands at approximately 920-966 EH/s, down from 1,066 EH/s in Q1 2026 — a 5.8% quarter-over-quarter decline. The Q1 decline itself was the first quarterly hashrate contraction in six years, breaking five consecutive years of double-digit annual growth, according to CoinDesk reporting.
Average block times extended to roughly 11.4 minutes during the February drawdown, above the 10-minute target, before the difficulty adjustment corrected toward equilibrium. The next adjustment is estimated for August 8, with CoinWarz projecting approximately 125.13 trillion, a further 0.87% decrease.
According to Hashrate Index data reported by Blockspace, this marks only the second year-over-year difficulty decline in Bitcoin's history. The first occurred during China's 2021 mining ban, which took roughly half the network's computing power offline in a matter of weeks.
The core driver is arithmetic. According to CoinShares' Q1 2026 Bitcoin Mining Report, the weighted average cash cost to produce one bitcoin among publicly listed miners rose to approximately $80,000 in Q4 2025. Bitcoin traded at approximately $62,800-$63,500 as of August 2-3, 2026. That implies losses of roughly $17,000-$19,000 per BTC mined for the average listed operator.
Hashprice, the industry-standard metric measuring expected miner revenue per petahash per second per day, fell to $27.66 in late June 2026 — a five-year low, according to CoinShares. It has since recovered modestly to $31.70 as of early August. Luxor's forward hashprice market prices an average of $31.85 per PH/s/day through December, suggesting market participants anticipate little improvement in mining economics for the remainder of 2026.
The economics bifurcate sharply by efficiency tier. Operations running sub-15 joules per terahash (J/TH) hardware with electricity costs at or below $0.04/kWh remain marginally profitable. Operators with older-generation rigs (above 20 J/TH) at residential electricity rates ($0.16-$0.20/kWh) are deep underwater. CoinShares estimates 15-20% of legacy mining rigs are now unprofitable at current hashprice levels. The firm projected further capitulation among higher-cost operators in H1 2026 unless BTC price recovers materially.
Publicly listed Bitcoin miners — including MARA Holdings, CleanSpark, Riot Platforms, Cango, Core Scientific, and Bitdeer — collectively sold more than 32,000 BTC in Q1 2026, according to data compiled by CoinShares. That figure exceeds their combined sales across all four quarters of 2025.
For context, the previous record for quarterly miner liquidation was approximately 20,000 BTC during Q2 2022, when the Terra-Luna collapse triggered a crypto-wide bear market. The Q1 2026 figure surpasses that by 60%.
The selling is driven by operational necessity. CleanSpark reported fiscal Q2 2026 revenue of $136.4 million, down 24.9% from $181.7 million a year earlier. Net loss widened to $378.3 million, including a $224 million BTC impairment charge. CleanSpark President and CFO Gary Vecchiarelli stated during the company's earnings call: "So we want to be redirecting every dollar possible toward AI capex."
The contrast with Bitcoin treasury companies is stark. Strategy (formerly MicroStrategy) has continued accumulating BTC throughout the same period. The divergence — miners selling to fund operations and AI transitions while treasury companies buy — represents a structural reallocation of Bitcoin ownership from producers to financial holders.
The migration of mining infrastructure toward AI and high-performance computing has accelerated beyond what most analysts anticipated. Publicly listed miners have committed to over $70 billion in cumulative AI/HPC contracts, according to industry tracking data.
The largest individual deals:
The economics justify the pivot. Nick Hansen, CEO and co-founder of Luxor mining pool, told DL News that "resisting the urge to transition to AI" will be miners' biggest challenge in 2026. Bernstein analysts noted that "Bitcoin miners are now an integral part of the AI value chain, providing warm powered shells for AI data centres." Beau Turner, CEO of Abundant Mines, told TheStreet Roundtable that companies pursuing AI strategies "have in many cases five to 10x the multiples on their forward-looking revenue" compared to pure-play mining operations.
The revenue differential is the deciding factor. As MARA Holdings CEO Fred Thiel stated plainly: revenue per unit of electricity consumed is substantially higher for AI workloads than for Bitcoin mining at current hashprice levels.
The hashrate decline is not solely an economics story. Regional disruptions have compounded the pressure.
Texas: ERCOT set new all-time electricity demand records in late July 2026, with preliminary peak demand hitting 91,308 MW on July 22. Bitcoin miners participating in voluntary curtailment agreements generated $56.7 million in demand-response credits across 2025 and $21 million in Q1 2026 alone — a 169% year-over-year increase. However, as miners convert facilities to AI data centers, the flexible curtailment capacity available to ERCOT diminishes. AI workloads cannot be interrupted without service degradation, unlike Bitcoin mining, which can resume seamlessly after power-downs.
Iran: The country's Bitcoin hashrate plunged from approximately 9 EH/s to 2 EH/s — a 77% decline — driven by ongoing crackdowns, conflict-related infrastructure damage, and energy disruptions. Iranian authorities report that 95% of the country's estimated 427,000 active mining devices operate illegally.
Broader factors: Hashrate Index attributed the 2026 network contraction to a combination of weak mining economics, AI/HPC capital reallocation, Texas summer power curtailments, Iran-related disruptions, and policy shocks across other mining jurisdictions.
The sustained hashrate decline raises questions about Bitcoin's proof-of-work security model, though the practical risk remains low by most assessments.
The cost to mount a theoretical 51% attack — procuring enough hardware to rival the network — still exceeds $10 billion in hardware alone, plus more than $1.5 million per hour in electricity, according to industry estimates. Mining pool concentration is a more immediate concern: Foundry USA controls roughly 30% of global hashpower, while AntPool, ViaBTC, and F2Pool collectively push the combined share above 70% during some windows.
Abundant Mines CEO Beau Turner argued that the public miner exodus does not fundamentally threaten the network: "I don't think this should concern Bitcoin holders at all. The security model of the network works really well regardless of whether the public miners are mining or not." Turner suggested that "mining is in the long run very unlikely to be in the hands of very large operators. It's going to naturally go toward the edges."
The difficulty adjustment mechanism — which recalibrates roughly every two weeks — continues to function as designed, ensuring block production returns to the 10-minute average target. The protocol's self-correcting nature means the network adapts to hashrate changes automatically, though prolonged declining difficulty does reduce the cumulative computational work securing the chain.
Canaan VP of Corporate Affairs Gwyn Lauber offered a historical perspective: "Margins are clearly under pressure right now, but Bitcoin mining has experienced many similar moments. An easing cycle in 2026 would likely result in better Bitcoin prices and mining margins."
The Bitcoin mining industry is undergoing a structural transformation, not a cyclical downturn. Prior difficulty declines — China's 2021 ban being the only precedent — were driven by acute regulatory events and resolved within months as hashrate relocated. The 2026 decline is driven by persistent economic fundamentals: post-halving revenue compression, rising energy costs, and a competing use case (AI infrastructure) that generates meaningfully higher revenue per unit of electricity.
The question is not whether marginal miners will capitulate — that is already occurring at scale. The question is whether the hashrate floor stabilizes at current levels or continues to erode as more operators convert remaining capacity to AI workloads. Luxor's forward market, pricing hashprice essentially flat through December, implies the market expects no resolution this year.
Bitcoin's protocol-level security model continues to function through the difficulty adjustment mechanism. The economic argument for proof-of-work, however, now competes directly with the economic argument for providing AI compute — a contest that, at current relative pricing, AI is winning decisively.