Public Bitcoin miners sold a record 32,000 BTC in Q1 2026 — more than the combined total for all of 2025, and exceeding the 20,000 BTC liquidated during the Q2 2022 Terra-Luna crash. The sell-off coincides with hashprice collapsing to $28–30 per petahash per second per day in early March, the low...
"We're not a Bitcoin mining company anymore. We're an energy infrastructure company that happens to mine Bitcoin on the side." — Paul Prager, CEO, TeraWulf, Q1 2026 Earnings Call
Public Bitcoin miners sold a record 32,000 BTC in Q1 2026 — more than the combined total for all of 2025, and exceeding the 20,000 BTC liquidated during the Q2 2022 Terra-Luna crash. The sell-off coincides with hashprice collapsing to $28–30 per petahash per second per day in early March, the lowest level since 2018, rendering 15–20% of the global mining fleet unprofitable according to CoinShares.
The capital is flowing into artificial intelligence infrastructure. TeraWulf, IREN, and Core Scientific have collectively locked in over $35 billion in contracted HPC revenue from hyperscalers including Microsoft, Google-backed Fluidstack, and CoreWeave. Mining stocks have outperformed Bitcoin by 70% year-to-date in 2026, with TeraWulf up 73.58% while BTC sits approximately 12% in the red since January 1. The market is repricing these firms not as miners but as data center operators with power purchase agreements worth multiples of their mining revenue.
Bitcoin's network hashrate has slipped below 1 zettahash per second for the first time since mid-2025, registering 977.89 EH/s as of early May. Mining difficulty has dropped in six of the nine adjustment epochs in 2026. The structural shift raises questions about network security concentration as public miners redirect capacity away from SHA-256 hashing.
Publicly listed Bitcoin miners — including Marathon Digital (MARA), CleanSpark (CLSK), Riot Platforms (RIOT), Cango, Core Scientific (CORZ), and Bitdeer — collectively offloaded more than 32,000 BTC during Q1 2026. The figure represents a new industry record for a single quarter, surpassing the previous high of 20,000 BTC sold during Q2 2022 when the Terra-Luna ecosystem collapse triggered broad liquidations across crypto.
The largest single disposition came from Marathon Digital, which sold 15,133 BTC between March 4 and March 25, tied to debt repurchases totaling approximately $1 billion. CleanSpark sold 553 of the 568 BTC it produced in February alone — nearly its entire monthly output. Riot Platforms offloaded 1,818 BTC (~$162 million) in December 2025, setting the pattern that accelerated into Q1.
The contrast with corporate Bitcoin treasury firms is stark. Strategy (formerly MicroStrategy) added 89,600 BTC for $5.5 billion in Q1 2026 — its second-largest quarterly purchase ever — bringing total holdings to 818,334 BTC valued at $64.44 billion. The mining industry is net selling while treasury firms are net buying, creating an unusual intra-industry flow dynamic.
Mining industry aggregate BTC reserves declined from over 1.86 million BTC at end-2023 to approximately 1.8 million BTC by April 2026. The direction is clear: miners are converting hashrate-generated BTC into fiat to fund capital-intensive data center buildouts.
The April 2024 halving reduced block rewards from 6.25 BTC to 3.125 BTC, cutting per-block mining revenue in half overnight. Eighteen months later, the compounding effects are acute.
Key profitability metrics (CoinShares Q1 2026 Mining Report):
| Metric | Value | Context | |--------|-------|---------| | Hashprice (March 2026 low) | $28–30/PH/s/day | Lowest since 2018 | | Hashprice (late April 2026) | $36.46/PH/s/day | +13.65% from March lows | | Weighted avg. cash cost per BTC | ~$79,995 | Among public miners, Q4 2025 | | MARA electricity cost per BTC | $64,703 | Diverse geography, third-party hosting | | CleanSpark electricity cost per BTC | $52,463 | Up from $44,679 in Q2 2025 | | IREN electricity cost per BTC | $34,325 | Favourable Childress, TX power agreements | | Unprofitable fleet share | 15–20% | At $30/PH/s/day hashprice | | Hash revenue per EH/s (Dec 2025) | $38,700/day | Down 32% year-over-year, record low |
At a hashprice of $30/PH/s/day, any miner running hardware below an S19 XP with electricity costs at or above $0.06/kWh is losing money. CoinShares estimates this covers 15–20% of the global mining fleet. The ROI on new ASIC hardware has stretched beyond 1,000 days at current economics, according to CCN analysis.
An extreme cold snap in January 2026 forced temporary shutdowns of approximately 35% of Bitcoin's hashrate across U.S. facilities. Many of those facilities never came back online for mining, instead being repurposed or permanently idled.
The transformation from Bitcoin miner to AI infrastructure provider is no longer theoretical. It is quantifiable in contracted revenue.
Major AI/HPC contracts signed by former miners:
| Company | Contract Partner | Contracted Revenue | Capacity | Duration | |---------|-----------------|-------------------|----------|----------| | IREN | Microsoft | $1.94B annualized | 4.5 GW pipeline, 140K NVIDIA GPUs | 5 years | | Core Scientific | CoreWeave | ~$10.2B | 590 MW critical IT load, 6 sites | 12 years | | TeraWulf | Google-backed Fluidstack, Core42 | $12.8B+ | 1+ GW available power | Multi-year | | Riot Platforms | AMD | 50 MW contracted | Doubled from 25 MW | Ongoing |
Total contracted AI/HPC revenue across leading miners exceeds $35 billion. Bloomberg reported in April 2026 that AI revenue is set to surpass Bitcoin mining revenue for leading crypto mining companies, with some operators on track to derive 70% of revenue from AI by year-end.
Revenue mix is shifting rapidly:
Core Scientific announced plans to repurpose its Pecos, Texas campus into a 1.5-gigawatt AI data center hub and is raising $3.3 billion through senior secured notes due 2031 to fund expansion across Georgia, Texas, North Carolina, and Oklahoma.
The market is pricing the AI pivot, not the mining economics. Most top-ten publicly listed mining operators have posted YTD gains of 25–73% while Bitcoin is approximately 12% negative since January 1.
YTD stock performance (as of early May 2026):
| Company | Ticker | YTD Performance | Primary Catalyst | |---------|--------|-----------------|------------------| | TeraWulf | WULF | +73.58% | $12.8B HPC contracts | | IREN | IREN | ~+60% | Microsoft partnership, $16.71B market cap | | Core Scientific | CORZ | ~+45% | CoreWeave $10.2B deal | | Riot Platforms | RIOT | ~+25% | Data center diversification | | Bitcoin | BTC | ~-12% | Post-January decline from $90K |
IREN now commands a $16.71 billion market capitalization — a valuation that reflects AI infrastructure expectations, not hashrate output. The collective mining sector has outperformed BTC by approximately 70% in 2026, per Bitcoin.com analysis.
S&P Global published research in February 2026 documenting the pivot, noting that Bitcoin miners are increasingly being covered by data center and infrastructure analysts rather than crypto-native research teams. The re-rating reflects a fundamental shift in how these companies generate free cash flow.
Bitcoin's network hashrate has fallen below 1 zettahash per second, registering 977.89 EH/s in early May 2026. This marks a decline from the peak of approximately 1.1 ZH/s in October 2025.
Mining difficulty has dropped in six of nine adjustment epochs in 2026. On May 1, difficulty eased 2.3% following a 2.43% decline on April 17. The total hashrate sits at approximately 920 EH/s by some estimates, well below the 1 ZH/s threshold that represented a milestone when first breached in 2025.
The cause is structural, not cyclical. Hashrate is leaving the network faster than fresh capacity is replacing it, because:
The declining difficulty does improve economics for remaining miners — hashprice rose 13.65% from March lows by late April. But CoinShares projects this window closes as sovereign miners and private operators fill the gap.
As public miners exit, nation-state and private operators are expanding. According to Hashrate Index, the United States retains 37.5% (~400 EH/s) of global hashrate as of January 2026, with Russia capturing 16.4% (~175 EH/s) benefiting from natural gas, hydropower, and Siberian cooling advantages.
Energy-rich nations including Turkmenistan and Bhutan are using state-owned investments to turn mining into a mechanism for monetizing surplus energy and accumulating strategic BTC reserves. The "sovereign miner" thesis — nations using mining as energy policy and reserve accumulation — is filling some of the gap left by public company departures.
CoinShares projects the network could rebound to 1.8 ZH/s by year-end 2026 as new capacity from sovereign operations, private miners, and next-generation ASIC deployments enters the network. However, the composition of miners will have shifted meaningfully from publicly-traded Western operators toward state-backed and private Asian/MENA entities.
The Bitcoin mining industry is undergoing an identity crisis that is, in economic terms, a rational reallocation of capital toward higher-margin activities. At $28–30/PH/s/day hashprice, mining Bitcoin generates roughly 8–12% gross margins for efficient operators and negative margins for the rest. AI colocation at 85% EBITDA margins (per IREN's Microsoft deal) offers a fundamentally different risk-return profile on the same power infrastructure.
The 32,000 BTC Q1 sell-off is not capitulation in the traditional sense — it is deliberate asset conversion from a depreciating revenue stream (BTC mining at sub-economic hashprice) into a growing one (contracted AI compute). The market agrees: mining stocks are up 70% versus BTC's 12% decline.
The open question is network security. If public miners continue redirecting gigawatts from SHA-256 to GPU compute, and sovereign/private miners fill the gap, Bitcoin's security model becomes more dependent on actors with less transparency and different incentive structures. CoinShares' projection of 1.8 ZH/s by year-end assumes new entrants materialize. If they don't, the current 920–977 EH/s hashrate level may represent a new equilibrium — one that is economically sustainable at current hashprice but represents a 10–15% decline from 2025 peaks in the cryptographic work securing the network.