Bitcoin's mining sector is undergoing its most severe profitability crisis since the April 2024 halving. Network hashrate has shed approximately 145 EH/s from its May 2026 peak above 1,000 EH/s, falling to 918 EH/s by mid-June. Difficulty dropped 10.09% on June 14 — the largest single adjustment ...
"There's going to be more and more demand for financial technologies, because there will be more and more transactions, so the industry will grow." — Changpeng Zhao, Founder, Binance
Bitcoin's mining sector is undergoing its most severe profitability crisis since the April 2024 halving. Network hashrate has shed approximately 145 EH/s from its May 2026 peak above 1,000 EH/s, falling to 918 EH/s by mid-June. Difficulty dropped 10.09% on June 14 — the largest single adjustment of 2026 and the 11th-biggest downward move in Bitcoin's history. Hashprice, the standard measure of mining revenue per unit of computational power, hit a post-halving record low near $27.89 per petahash per day, down 66% from its October 2025 peak.
The squeeze is structural. According to CoinShares' Q1 2026 mining report, the weighted average cash cost to produce one bitcoin among publicly listed miners rose to approximately $79,995 in Q4 2025. With BTC trading near $60,000 through late June 2026, an estimated 20% of legacy hashrate is now operating below breakeven. Public miners have responded by liquidating more than 15,000 BTC from peak treasury levels and accelerating a strategic pivot toward AI data center infrastructure, where contracted revenue now exceeds $13 billion across multi-year leases.
This report examines the mechanics of the current capitulation, quantifies the profitability breakdown across hardware tiers, and assesses whether the mining industry's AI pivot represents a sustainable economic transformation or a liquidity bridge to the next cycle.
Bitcoin's network hashrate peaked above 1,000 EH/s in early 2026 before entering a sustained decline. By early June, hashrate had fallen to approximately 893 EH/s — a drop of more than 100 EH/s. The seven-day moving average settled near 918 EH/s by mid-month, according to data from CoinWarz and Hashrate Index.
The decline triggered a 10.09% downward difficulty adjustment at block height 953,568 on June 14, 2026. Difficulty fell to 124.93 trillion, its lowest level since July 2025. The adjustment ranks as the second-largest negative retarget of 2026 and the 11th-biggest in the network's history, according to Bitcoin Foundation data.
Not all of the hashrate reduction represents permanent shutdown. Industry analysts note that Texas-based miners routinely curtail operations during summer peak demand periods under demand-response agreements with grid operators like ERCOT. This seasonal curtailment accounts for a portion of the observed decline, though the magnitude of the drop — roughly 145 EH/s from peak — exceeds typical seasonal patterns.
For miners still operating, the difficulty reduction delivers approximately 10% more BTC per unit of hash per day. This mechanical relief provides a narrow reprieve, but it does not resolve the fundamental mismatch between production costs and market price.
Hashprice — the dollar-denominated revenue a miner earns per petahash per second per day — fell to approximately $27.89/PH/s/day in early June 2026. This represents a 66% decline from the October 2025 peak and a new all-time post-halving low, according to Hashrate Index data.
The metric briefly recovered to approximately $32.31/PH/s/day following the June 14 difficulty adjustment, reflecting the mechanical increase in BTC earned per unit of hash. However, this remains well below levels required for breakeven operations on mid-generation hardware.
May 2026 aggregate miner revenue was approximately $1.12 billion, down 26% year-over-year, according to on-chain data. The revenue decline reflects the compounding effect of lower BTC prices (down approximately 21% from early May levels of $81,000 to $63,865 by mid-June), the April 2024 halving's ongoing reduction in block subsidy, and transaction fee revenue that has not compensated for the subsidy reduction.
CoinShares' Q1 2026 Bitcoin Mining Report documented a weighted average cash cost of approximately $79,995 per BTC among publicly listed miners as of Q4 2025. Conditions deteriorated further into Q1 2026. With BTC trading in the $58,000–$63,000 range through late June, miners face an estimated loss of approximately $17,000–$22,000 per BTC mined at industry-average cost structures.
Profitability thresholds vary significantly by hardware generation and electricity cost:
| Hardware Class | Efficiency (J/TH) | Breakeven Electricity Rate (at ~$60K BTC) | |---|---|---| | Latest-gen (S21-class) | <15 J/TH | ~$0.10/kWh | | Mid-gen (S19j Pro-class) | ~29.5 J/TH | ~$0.04/kWh | | Legacy (S19-class) | ~34+ J/TH | Unprofitable at most rates |
According to CoinShares, 15%–20% of legacy mining rigs are now operating below their electricity-only breakeven point. For most mid-tier operators, the electricity-only breakeven BTC price sits near $74,000, while the all-in operating cash flow breakeven is approximately $77,000.
Electricity costs account for 60%–80% of total mining operating expenses. The difference between $0.04/kWh and $0.08/kWh determines whether mid-generation hardware generates positive or negative margins. At current difficulty and BTC prices, only operations running latest-generation ASICs at sub-$0.05/kWh power rates maintain meaningful positive cash flow.
The three largest publicly traded Bitcoin miners — MARA Holdings, CleanSpark, and Riot Platforms — all reported declining June production:
| Company | June 2026 Production | MoM Change | BTC Holdings (June 30) | June Sales | |---|---|---|---|---| | MARA Holdings | 713 BTC | -25% | 49,940 BTC | 0 BTC | | CleanSpark | 685 BTC | -1.3% | 12,608 BTC | 578.51 BTC | | Riot Platforms | 450 BTC | -12.4% | 19,273 BTC | 397 BTC |
The production declines reflect both lower hashrate allocation and the mechanical impact of difficulty levels prior to the June 14 adjustment.
More significantly, MARA has shifted its treasury policy. The company liquidated approximately 20,880 BTC in Q1 2026, yielding roughly $1.5 billion at an average realization of $70,137 per BTC. Year-to-date, MARA's net BTC position has declined by approximately 17,947 BTC — a 33.7% reduction from peak levels. The company disclosed in a filing that it expanded its crypto management strategy to permit balance sheet BTC sales, reversing its prior accumulation stance.
Core Scientific sold $175 million worth of Bitcoin — nearly 2,000 coins — in March 2026 to fund its infrastructure transition. The company was subsequently acquired by CoreWeave for $9 billion in stock.
According to CoinShares, public miners have collectively reduced BTC treasuries by more than 15,000 BTC from peak levels.
The mining industry's response to the profitability crisis extends beyond cost-cutting. A structural pivot toward AI and high-performance computing (HPC) infrastructure is underway, with multiple public miners targeting 70% of revenue from AI by end of 2026.
Key transactions and contracts:
The economic logic is straightforward: AI data center hosting offers higher revenue per kilowatt-hour than Bitcoin mining during periods of compressed hashprice, with more predictable cash flows via long-term contracts. Both operations share cooling, power, and physical infrastructure requirements.
The pivot has been reflected in equity markets. All ten of the largest publicly traded mining stocks were positive year-to-date through June 2026, with gains ranging from approximately 5% to 85%, according to industry trackers. This divergence from BTC's approximate 50% decline from its October 2025 all-time high above $126,000 indicates that markets are repricing these companies as AI infrastructure providers rather than pure Bitcoin mining operations.
However, the transition carries execution risk. GPU-based AI workloads require different cooling, power density, and networking infrastructure than SHA-256 ASIC mining. Companies must invest significant capital to retrofit facilities, and AI demand cycles — while currently robust — are not guaranteed over the 10–25 year lease horizons being contracted.
CryptoQuant's Hash Ribbons indicator — which tracks the crossover between 30-day and 60-day hashrate moving averages — signaled miner capitulation in late May/early June 2026. Historically, this signal has preceded price recoveries. The most recent comparable signal occurred in July 2025, which preceded an approximately 25% price rally over the following months.
The capitulation mechanism functions as a sorting process: higher-cost operators exit, difficulty adjusts downward, and surviving low-cost miners capture a larger share of block rewards at improved margins. This cycle has repeated in every prior Bitcoin bear market.
However, the current cycle differs in one material respect. Previous capitulations resolved primarily through BTC price recovery. This cycle features a parallel exit pathway via AI infrastructure conversion, meaning that some hashrate may not return to Bitcoin mining even if prices recover. The long-term implications for network hashrate concentration and decentralization remain uncertain.
The Bitcoin mining industry is in the deepest profitability crisis of the post-halving era. The confluence of a 50% BTC price decline from all-time highs, the halving's sustained subsidy reduction, and rising production costs has pushed an estimated one-fifth of network hashrate below breakeven.
The industry's response is bifurcated. Low-cost operators with latest-generation hardware and sub-$0.05/kWh power access continue to mine profitably and benefit from each difficulty reduction. High-cost operators face a binary choice: shut down mining operations or convert infrastructure to AI hosting, where long-term contracts offer more predictable economics.
This structural transformation has material implications for Bitcoin's network security model. If AI hosting permanently absorbs mining infrastructure, the network may emerge from this cycle with fewer, larger mining operators concentrated in the lowest-cost power markets. The economic value distribution within the mining sector — historically broad — is narrowing toward a winner-take-most structure.
The data does not yet support a definitive conclusion on whether the AI pivot represents permanent hashrate loss or a temporary capital bridge. What is clear: the mining industry that emerges from this capitulation will bear little resemblance to the one that entered it.