Publicly listed Bitcoin miners sold over 32,000 BTC in Q1 2026 — more than their combined net sales for all of 2025 and more than the approximately 20,000 BTC liquidated during the Terra-Luna collapse in Q2 2022. The sell-off, led by MARA Holdings ($1.1 billion), Riot Platforms ($289.5 million), ...
"Revenue diversification has become the definitive line between life and death for miners." — CoinShares Research, Q1 2026 Bitcoin Mining Report
Publicly listed Bitcoin miners sold over 32,000 BTC in Q1 2026 — more than their combined net sales for all of 2025 and more than the approximately 20,000 BTC liquidated during the Terra-Luna collapse in Q2 2022. The sell-off, led by MARA Holdings ($1.1 billion), Riot Platforms ($289.5 million), and Core Scientific ($175 million in January alone), coincided with hashprice falling to $28–30 per PH/s/day, a post-halving all-time low.
The liquidation is financing an industry-wide pivot. Bitcoin miners have announced over $70 billion in cumulative AI and high-performance computing (HPC) contracts with major technology companies. CoinShares projects that listed miners could derive 70% of their revenues from AI by December 2026, up from approximately 30% today. The economic logic is straightforward: AI colocation generates roughly three times the revenue per megawatt compared to Bitcoin mining.
The consequences for Bitcoin's network security are measurable. Hashrate fell from a peak of 1,160 EH/s in October 2025 to approximately 1,020 EH/s in Q1 2026 — a 12% decline. Two mining pools, Foundry USA and AntPool, now control over 51% of the remaining hashrate. An estimated 85% of global hashrate is projected to be controlled by just 12 publicly traded entities or sovereign wealth funds by year-end 2026. Bitcoin's decentralization assumption is under structural pressure.
The scale of Q1 2026 miner selling was without recent precedent. According to data compiled by The Block and confirmed across multiple sources, publicly traded mining companies collectively sold more than 32,000 BTC during the quarter. For context, total net sales across the same cohort for all four quarters of 2025 were lower than this single quarter.
Breakdown of major sellers (Q1 2026):
| Company | BTC Sold | Approximate Proceeds | Notes | |---------|----------|---------------------|-------| | MARA Holdings | 15,133 | $1.1 billion | Sold between March 4–25 to retire ~$1B convertible notes | | Riot Platforms | 3,778 | $289.5 million | Follow-on to MARA's exit | | Core Scientific | ~1,900 | ~$175 million | January sales to fund AI conversion | | CleanSpark | ~553 (Feb) | N/A | Sold 553 of 568 BTC mined in February | | Bitdeer | Undisclosed | N/A | Produced 668 BTC in January; partial sales | | Cango | Undisclosed | N/A | Net seller throughout Q1 |
MARA's sale was the single largest corporate Bitcoin liquidation since the strategy of holding BTC on balance sheets became industry standard. The company used $1 billion in proceeds to repurchase convertible senior notes, cutting debt by approximately 30%. MARA still holds 53,822 BTC, according to company disclosures as of April 2026.
CleanSpark's February production data illustrates the severity: the company mined 568 BTC and sold 553, retaining just 15 BTC — a 97% sell-through rate that signals operational treasury depletion rather than strategic selling.
The 2024 halving reduced block rewards from 6.25 BTC to 3.125 BTC. Eighteen months later, the full economic impact is visible in cost structures that have outrun revenue for much of the industry.
All-in cost per BTC mined (Q4 2025, latest available):
| Company | All-In Cost/BTC | Cash Cost/BTC | |---------|----------------|---------------| | CleanSpark (CLSK) | $118,932 | Lowest peer SG&A at $17,848/BTC | | IREN | $140,441 | Electricity cost: $52,463/BTC | | MARA Holdings | $153,040 | Highest production volume: 2,011 BTC/quarter | | Hut 8 | $160,402 | $411M total debt | | Core Scientific | $168,693 | 39% revenue already from AI colocation | | Riot Platforms | $170,366 | Sold 3,778 BTC in Q1 | | HIVE Digital | $144,321 | 5% HPC revenue mix | | Bitdeer | $118,188 | Lowest all-in cost; 70 EH/s self-mining |
The weighted average cash cost across public miners reached approximately $79,995 per BTC in Q4 2025, according to CoinShares. With Bitcoin trading in the $68,000–$75,000 range for portions of Q1 2026, a significant share of the industry was mining at a loss.
Hashprice — the standard metric for miner revenue per unit of computation — tells the same story. From a Q3 2025 peak of approximately $63/PH/s/day, hashprice collapsed to $28–30/PH/s/day by early March 2026, reaching what CoinShares described as a post-halving all-time low. A difficulty adjustment on April 17 provided partial relief, with hashprice recovering to $36/PH/s/day, but this remains well below levels most operators require for sustained profitability.
An estimated 15–20% of the global mining fleet is currently operating at a loss, according to the CoinShares Q1 2026 report. The breakeven electricity price for Bitmain's S19 XP — still a widely deployed unit — fell to $0.077/kWh by December 2025, meaning any operator paying above that rate is underwater on that hardware.
The migration from Bitcoin mining to AI infrastructure is no longer speculative. It is underwritten by signed contracts worth tens of billions of dollars.
Major AI/HPC contracts secured by former or current Bitcoin miners:
| Company | Contract Partner | Value | Duration | Capacity | |---------|-----------------|-------|----------|----------| | Core Scientific | CoreWeave | $10.2 billion | 12 years | N/A | | IREN | Microsoft | $9.7 billion | N/A | 76,000 NVIDIA GB300 GPUs, 200 MW (Childress, TX) | | Hut 8 | Fluidstack | $7 billion | 15 years | 245 MW | | WULF (TeraWulf) | Various | $12.8 billion | Contracted HPC pipeline | N/A |
Cumulative announced AI/HPC contracts by Bitcoin miners exceeded $70 billion as of Q1 2026. The economic rationale is quantifiable: AI colocation generates approximately three times the revenue per megawatt compared to Bitcoin mining, according to CoinShares data.
Current AI revenue contribution (Q4 2025):
CoinShares projects that mining revenue will plummet from approximately 85% of total revenue in early 2025 to less than 20% by end-2026 for companies that have secured AI contracts. In effect, the publicly traded "Bitcoin mining" sector is becoming an AI infrastructure sector that happens to mine Bitcoin as a secondary activity.
The market has priced this shift. As of January 2026, AI-pivoted mining firms trade at nearly double the valuation per megawatt of power capacity compared to Bitcoin-focused peers, according to financial analysis cited in multiple reports.
The migration of hashrate from Bitcoin to AI compute has measurable security implications.
Hashrate trajectory:
The network experienced three consecutive negative difficulty adjustments in late 2025, the first such streak since July 2022. Difficulty fell 2.43% on April 17, 2026, to 135.59 trillion.
Transaction fees — the long-term replacement for block subsidy as the network's security budget — stabilized at approximately 15% of miner revenue in 2026. Industry consensus holds that fees must consistently exceed 20% of miner revenue for the network to maintain adequate security incentives without block rewards. The current 15% ratio falls short of that threshold, though the next halving is not expected until approximately 2028.
The CoinShares year-end projection of 1.8 ZH/s depends on institutional buildout continuing. If Bitcoin prices remain in the $70,000–$80,000 range while AI revenue opportunity persists, the economic incentive to redirect power capacity away from mining remains strong.
The departure of marginal miners has concentrated hashrate among fewer, larger operators — and their hashrate is further concentrated into a small number of mining pools.
Mining pool concentration (2026):
| Pool | Hashrate Share | Approximate Hashrate | |------|---------------|---------------------| | Foundry USA | ~37% | ~277 EH/s | | AntPool (Bitmain) | ~14% | ~146 EH/s | | ViaBTC | ~12% | ~120 EH/s | | F2Pool | ~8% | ~77 EH/s | | Binance Pool | ~5% | ~54 EH/s |
Foundry USA and AntPool together control over 51% of blocks mined. Pool operators — not individual miners — build block templates and determine transaction selection. Individual miners contributing hashrate to these pools have no direct say in which transactions are included or excluded from blocks.
The hardware supply chain compounds the risk. Three ASIC manufacturers — Bitmain, MicroBT, and Canaan — dominate production. Supply chain disruption or regulatory intervention at the hardware level would cascade into mining capacity, pool distribution, and ultimately protocol-level operations.
A 2025 stress test demonstrated the fragility of this concentration: a severe winter storm in the United States caused a 60% drop in Foundry USA's hashrate, with approximately 200 EH/s going offline. A single weather event temporarily removed roughly 20% of the network's total hashrate from a single pool.
The United States, China, and Russia now control approximately 68% of global hashrate, according to CoinShares. The U.S. gained approximately 2 percentage points of market share in Q4 2025.
Emerging mining jurisdictions include Paraguay, Ethiopia, and Oman, which have entered the global top-10 mining countries. Bhutan's sovereign wealth fund, Druk Holding and Investments, which partnered with Bitdeer on a $500 million carbon-neutral mining fund, has reduced its Bitcoin holdings from approximately 13,000 BTC in late 2024 to approximately 5,400 BTC as of March 2026 — selling approximately $42.5 million worth of BTC and USDT in 2026.
The geographic pattern mirrors the corporate one: sovereign and quasi-sovereign entities are entering mining while smaller private operators exit, further concentrating control among entities with access to cheap power and patient capital.
Record liquidation: Public miners sold 32,000+ BTC in Q1 2026, exceeding full-year 2025 sales and the Terra-Luna era liquidation, driven by hashprice at a post-halving all-time low of $28–30/PH/s/day.
Cost crisis: Weighted average cash cost to mine one BTC reached ~$80,000 in Q4 2025. All-in costs for major miners ranged from $118,188 (Bitdeer) to $170,366 (Riot). An estimated 15–20% of the global fleet is mining at a loss.
AI contracts total $70B+: Core Scientific ($10.2B with CoreWeave), IREN ($9.7B with Microsoft), Hut 8 ($7B with Fluidstack), and TeraWulf ($12.8B pipeline) have signed binding AI/HPC deals that dwarf mining revenue.
Revenue inversion approaching: CoinShares projects mining will drop from 85% to under 20% of listed miner revenue by end-2026, with AI filling the gap. The "Bitcoin mining" sector is being repriced as AI infrastructure.
Network centralization intensifying: Foundry USA and AntPool control 51%+ of hashrate. An estimated 85% of global hashrate will be controlled by 12 entities by year-end. Three ASIC manufacturers control hardware supply.
Security budget gap: Transaction fees at 15% of miner revenue remain below the 20% threshold considered necessary for long-term network security without block rewards.
The Q1 2026 data describes an industry undergoing structural transformation, not a cyclical downturn. The halving cut revenue in half; the AI opportunity provides an exit. The 32,000 BTC liquidation is the mechanism by which Bitcoin mining companies are financing their transition into AI infrastructure companies.
The network-level consequences are not theoretical. Hashrate has declined 12% from its October 2025 peak. Two pools control majority hashrate. The security budget remains below the threshold that researchers consider sustainable. Each miner that redirects power capacity from Bitcoin to AI reduces the pool of independent hashrate contributors, increasing the relative weight of the remaining large operators.
CoinShares' projection of 1.8 ZH/s by year-end 2026 requires sustained institutional investment in mining capacity. If the AI revenue premium persists — and $70 billion in signed contracts suggests it will — the rational economic decision for operators with dual-use infrastructure is to allocate capacity to the higher-returning use case. Bitcoin mining is becoming the residual activity performed with leftover capacity, not the core business.
The question for the network is whether a smaller, more concentrated set of miners can provide adequate security. The data does not yet answer that question definitively, but the trend lines are clear.