Bitcoin's mining sector is undergoing the most severe structural realignment in its history. The network's hashrate posted its first quarterly decline in six years during Q1 2026, falling approximately 4% year-to-date. Publicly listed miners now lose an estimated $19,000 on every bitcoin they pro...
"In time, we will have no bitcoin." — Ben Gagnon, CEO, Bitfarms (now Keel Infrastructure)
Bitcoin's mining sector is undergoing the most severe structural realignment in its history. The network's hashrate posted its first quarterly decline in six years during Q1 2026, falling approximately 4% year-to-date. Publicly listed miners now lose an estimated $19,000 on every bitcoin they produce, with weighted average cash costs near $80,000 per BTC against a spot price hovering around $67,000-$68,000. The next difficulty adjustment on April 19 is projected to drop 14.27% — the largest single reduction since 2021.
The response has been uniform and decisive: miners are exiting. Over $70 billion in AI and high-performance computing (HPC) contracts have been signed by public mining companies. Bitfarms has rebranded to Keel Infrastructure and declared it will hold zero bitcoin. MARA Holdings sold 15,133 BTC for $1.1 billion in March to retire debt and fund AI capacity. CoinShares projects some miners could derive up to 70% of total revenue from AI hosting by year-end 2026.
The economic logic is clear — AI contracts generate roughly three times the revenue per megawatt compared to bitcoin mining, with 80-90% operating margins. But the migration carries a second-order consequence for Bitcoin itself: each miner that exits reduces independent hashrate contributors, concentrating power among fewer pools and raising network centralization risk.
The April 2024 halving cut block rewards from 6.25 to 3.125 BTC. Twenty-four months later, the sector has not recovered.
According to CoinShares' Q1 2026 Bitcoin Mining Report, the weighted average cash cost to produce one bitcoin among publicly listed miners stands at approximately $79,995. With BTC trading in the $67,000-$70,000 range, the median public miner operates at a loss of roughly $10,000-$19,000 per coin produced. Transaction fees provide no cushion: fee revenue represents barely 0.56% of the block reward.
Hash price — the standard metric for miner revenue per unit of computational power — collapsed to approximately $28-30 per petahash per second (PH/s) per day in early Q1 2026, a post-halving all-time low. At this level, according to CoinShares, any miner running hardware below an Antminer S19 XP with electricity costs at or above $0.06/kWh is unprofitable. That threshold covers an estimated 15-20% of the global mining fleet.
The math is simple. At $30/PH/s/day and 3.125 BTC per block, miners need BTC above approximately $90,000 for most operations to break even on a fully loaded cost basis. The market has not cooperated.
Global hashrate peaked above 1.1 zettahash per second (ZH/s) — equivalent to 1,100 exahash per second (EH/s) — in late 2025. By Q1 2026, it had fallen to approximately 1,004 EH/s, according to data from CoinWarz and Hashrate Index. This marks the first first-quarter decline since 2020, ending five consecutive years of double-digit growth.
The decline accelerated in April 2026. The network currently produces blocks at an average interval of 11 minutes 39 seconds, well above the 10-minute target. This slowdown confirms sustained hashrate withdrawal — miners are shutting down machines faster than new capacity comes online.
A stress test occurred in January 2026 when Winter Storm Fern knocked out power across much of the United States. Hashrate temporarily plunged from ~1.1 ZH/s to 663 EH/s over a single weekend — a 30-40% drop — exposing the geographic concentration of mining capacity in the U.S., which accounts for 37.4% of global hashrate.
Three countries — the United States (37.4%), Russia (16.9%), and China (12.0%) — control roughly 65% of total network computational power. At the pool level, the top two mining pools held 55% of global hashrate as of late 2023, according to b10c's Mining Centralization Index, and concentration has worsened since.
The scale of capital reallocation from bitcoin mining to AI infrastructure is without precedent in the sector.
| Company | AI/HPC Contract Value | Contract Term | Capacity | |---------|----------------------|---------------|----------| | Core Scientific / CoreWeave | $10.2 billion | 12 years | ~590 MW | | TeraWulf | $12.8 billion | Multi-year | 510 MW | | Hut 8 (River Bend) | $7 billion | 15 years | Not disclosed | | Applied Digital / CoreWeave | $11 billion | Multi-year | Not disclosed |
Aggregate AI and HPC contracts across public mining companies exceed $70 billion. CoinShares projects mining revenue could fall from approximately 85% of total revenue in early 2025 to less than 20% by end of 2026 for companies with AI contracts in place.
The economics explain the migration. AI hosting contracts generate roughly three times the revenue per megawatt compared to bitcoin mining. Operating margins on AI deals run at 80-90%, compared to the negative margins most miners currently face on BTC production. Miners with secured HPC contracts now trade at 12.3x next-twelve-month sales, versus 5.9x for pure-play miners.
Bitfarms / Keel Infrastructure. The most definitive exit. Shareholders approved a U.S. re-domiciliation and rebrand to Keel Infrastructure (ticker: KEEL) effective approximately April 1, 2026. CEO Ben Gagnon stated the company will sell bitcoin "opportunistically into strength" and ultimately hold zero BTC. Bitfarms is advancing a 2.2 GW development pipeline targeting AI-driven revenue beginning in 2027. Its Washington State facility is being retrofitted with Nvidia GB300 GPUs and liquid cooling.
MARA Holdings. Between March 4 and March 25, 2026, MARA sold 15,133 BTC for approximately $1.1 billion. The company used roughly $1.0 billion to repurchase convertible notes at a ~9% discount, capturing approximately $88 million in value. MARA's February 2026 alliance with Starwood Capital targets conversion of mining sites into AI compute campuses with a path toward 2.5 GW of IT capacity. MARA still holds approximately 34,818 BTC following the sale.
CleanSpark. Operates as one of the remaining pure-play miners, maintaining fleet efficiency of ~16 W/T at approximately 50 EH/s installed capacity. SG&A costs of $17,848/BTC and stock-based compensation of $6,662/BTC are among the lowest in the peer group. CleanSpark has not announced AI contracts, making it a test case for the viability of pure mining.
Core Scientific. The company that arguably started the AI pivot trend with its CoreWeave partnership now has ~590 MW of contracted HPC capacity and 185+ MW of active billable capacity. Q4 2025 filings detail a $10 billion+ AI build-out plan.
The economic value analysis of Bitcoin has always depended on a critical assumption: sufficient independent hashrate to secure the network against attack. The current miner exodus tests that assumption.
Each miner that exits bitcoin for AI compute reduces the pool of independent hashrate contributors. This increases the relative weight of remaining large pools, worsens the Mining Centralization Index, and raises the probability of protocol-level stress events.
ASIC production remains dominated by three firms — Bitmain, MicroBT, and Canaan. Supply chain disruptions or regulatory action at the hardware level cascade directly into mining capacity. As the number of economically viable miners shrinks, the network becomes more dependent on fewer participants who remain profitable — primarily those with the lowest electricity costs and newest hardware.
The January 2026 Winter Storm Fern incident demonstrated the fragility: a single weather event in one country temporarily removed 30-40% of global hashrate. If miner exits continue at the current pace, the network becomes more vulnerable to similar concentrated disruptions.
It is worth noting that Bitcoin's difficulty adjustment mechanism is designed precisely for this scenario — difficulty drops, margins improve for remaining miners, and equilibrium reasserts itself. The question is whether the transition period introduces unacceptable concentration risk.
The projected 14.27% difficulty reduction on April 19, 2026 is the market's built-in stabilizer. By reducing the computational work required to mine a block, the adjustment immediately improves margins for surviving miners.
At current BTC prices of ~$67,000 and post-adjustment difficulty, the breakeven cost for efficient operations running latest-generation ASICs at $0.04/kWh drops to approximately $55,000-$60,000 per BTC — restoring positive margins for the lowest-cost quartile.
The difficulty adjustment also slows BTC selling pressure. Miners in margin distress sell production immediately and draw down reserves to fund operations. If margins improve post-adjustment, forced selling declines. Listed mining companies have collectively reduced BTC holdings by more than 15,000 BTC from their peak in recent months.
However, the adjustment does not reverse the AI migration. Miners are not leaving because of a temporary margin squeeze. They are leaving because AI hosting offers structurally superior unit economics. The difficulty adjustment can restore mining profitability at the margin, but it cannot make mining more profitable than AI hosting.
The bitcoin mining sector is repricing in real time. The April 2024 halving created the margin pressure; the AI infrastructure boom provided the exit. What remains is a smaller, more concentrated mining network that Bitcoin's difficulty adjustment will attempt to stabilize.
The economic question is whether Bitcoin at $67,000 can sustain a mining ecosystem large enough to maintain network security and decentralization. The difficulty mechanism suggests it can — at a lower hashrate, with fewer participants, operating at tighter margins. Whether that configuration is sufficient depends on how far the AI migration extends and how concentrated the remaining hashrate becomes.
For the mining companies themselves, the calculus is settled. Three times the revenue per megawatt with 80-90% margins versus negative margins on BTC production is not a difficult decision. The sector is no longer asking whether to pivot. It is asking how fast.