Public Bitcoin miners posted $4.3 billion in combined Q1 2026 net losses and liquidated over 32,000 BTC — a single-quarter record — as hashprice collapsed to $23.90/PH/s/day, its lowest level since 2018. The average network hashrate fell from 985 EH/s to 873 EH/s quarter-over-quarter, the first Q...
"Bitcoin miners have energy available today. It's an easy pivot. The key is being able to do it with the right partner." — Fred Thiel, CEO, MARA Holdings
Public Bitcoin miners posted $4.3 billion in combined Q1 2026 net losses and liquidated over 32,000 BTC — a single-quarter record — as hashprice collapsed to $23.90/PH/s/day, its lowest level since 2018. The average network hashrate fell from 985 EH/s to 873 EH/s quarter-over-quarter, the first Q1 decline in six years and the steepest drawdown since China's 2021 mining ban. Three consecutive negative difficulty adjustments, the first such streak since July 2022, confirmed miner capitulation.
The cause is structural, not cyclical. With all-in production costs near $90,000 per bitcoin and spot prices averaging $67,000–$68,000 through Q1, pure-play mining margins turned negative. Simultaneously, hyperscaler demand for GPU-ready data center capacity has handed the same companies an exit ramp: over $70 billion in aggregate AI/HPC hosting contracts have been signed across the public mining sector. TeraWulf's HPC revenue exceeded its Bitcoin mining revenue for the first time in Q1. Core Scientific holds a $10.2 billion, 12-year contract with CoreWeave. MARA sold $1.5 billion worth of BTC to retire 30% of its convertible debt and fund a $1.5 billion gas-fired data center acquisition.
The sector is bifurcating. Miners with grid-connected, low-cost power assets and hyperscaler contracts — Core Scientific, TeraWulf, Riot — saw their stocks rise 46–85% year-to-date even as they reported billion-dollar losses. Those without AI pipelines face insolvency risk. The implications for Bitcoin's security model — fewer ASICs, more geographic concentration, declining difficulty — remain unresolved.
The five largest publicly listed Bitcoin miners by market cap reported the following Q1 2026 results:
| Company | Revenue | YoY Change | Net Loss | BTC Sold | |---------|---------|------------|----------|----------| | MARA Holdings | $174.6M | -18.3% | -$1.3B | 20,880 BTC | | Riot Platforms | $167.2M | — | -$1.44/share | — | | CleanSpark | $136.4M | -24.9% | -$378.3M | — | | Hut 8 | — | — | -$253.1M | — | | Bitdeer | — | — | -$39M (gross) | — |
According to CoinShares' Q1 2026 Bitcoin Mining Report, revenue declines were driven by two compounding factors: Bitcoin's 22% price decline from approximately $87,500 to $68,200 during the quarter, and rising network difficulty that compressed per-unit economics. Bitdeer reported a gross margin of negative 20.7%, absorbing $70 million in noncash mining rig depreciation alongside elevated seasonal power costs.
American Bitcoin Corp. (ABTC), the Trump-family-linked mining venture, reported an $82 million Q1 net loss, widening from Q4 2025 due to the Bitcoin price decline and a $117 million book impairment. The company maintained a gross margin of 52% through operational efficiencies, according to its earnings filing.
Hashprice — the expected daily revenue per petahash per second of deployed mining power — is the core viability metric for the industry. It peaked at approximately $63/PH/s/day in July 2025, declined to $36–$38/PH/s/day by Q4 2025, and collapsed to $23.90/PH/s/day by the end of Q1 2026, according to CoinShares. That reading is the lowest since 2018.
The decline triggered measurable capitulation:
CoinShares projects that if Bitcoin recovers to $100,000, hashprice would rebound to approximately $37/PH/s/day. If prices remain below $80,000, continued difficulty increases would push hashprice lower — unless enough miners shut down to stabilize it through reduced competition.
Publicly listed miners collectively sold more than 32,000 BTC in Q1 2026, according to data compiled by BeInCrypto, exceeding the total BTC sold by public miners across all of 2025. The liquidation was driven by negative operating margins and the capital requirements of the AI infrastructure buildout.
MARA Holdings accounted for the largest single-company sell-off: 20,880 BTC ($1.5 billion), including a $1.1 billion sale near quarter-end to fund a convertible note repurchase. The company used $1 billion from the proceeds to retire 30% of its convertible debt, reducing total leverage from $3.3 billion to $2.3 billion. Despite the liquidation, MARA remains the fourth-largest corporate Bitcoin holder with 35,303 BTC ($2.84 billion at current prices).
Core Scientific sold approximately 1,900 BTC ($175 million) in January and disclosed plans to liquidate substantially all remaining BTC holdings by end of 2026, according to CoinShares.
The sell-off has implications beyond miner balance sheets. According to CoinDesk, the aggregate BTC selling pressure from public miners created measurable downward force during Q1, contributing to the 22% price decline. Corporate Bitcoin holdings across all companies stood at 1.15 million BTC at the end of Q1, according to Cryptopolitan — meaning miners' 32,000 BTC liquidation represented approximately 2.8% of the total corporate stack.
The Bitcoin mining-to-AI infrastructure transition, which began as a narrative in late 2024, is now generating measurable revenue. Over $70 billion in aggregate AI and HPC hosting contracts have been announced across the public mining sector, according to CoinShares.
Core Scientific (CORZ): Holds a 12-year, $10.2 billion agreement with CoreWeave for approximately 590 MW of HPC capacity. Total pipeline stands at 3.0 GW including 520 MW of existing capacity, 285 MW from the Hunt County (TX) acquisition, 330 MW from the Muskogee (OK) acquisition, and additional expansion opportunities.
TeraWulf (WULF): Secured $12.8 billion in HPC contracts, according to Bitcoin.com. In Q1 2026, HPC revenue exceeded Bitcoin mining revenue for the first time, with 60 MW of operational HPC capacity at its Lake Mariner data center. Plans call for additional expansion through 2026.
Riot Platforms (RIOT): Signed an initial 25 MW lease with AMD, subsequently expanded to 50 MW with potential to reach 150 MW, generating an estimated $636 million over a 10-year term, according to CoinDesk. Separately, Riot announced a collaboration with Terrestrial Energy to develop nuclear-powered data centers with up to 4 GW of nuclear capacity, with deployment expected in the early 2030s.
MARA Holdings (MARA): Partnered with Starwood Capital and agreed to acquire Long Ridge Energy & Power, a gas-fired power plant and data center in Ohio, for $1.5 billion. The site could support more than 600 MW of AI load. CEO Fred Thiel stated the company would not pursue large-scale ASIC purchases going forward, according to its Q1 filing.
Industry projections cited by CoinTelegraph suggest AI and HPC could account for 70% of revenue for transformed miners by year-end 2026, representing a potential $40 billion revenue opportunity across the sector.
Despite aggregate Q1 losses exceeding $4 billion, publicly listed mining stocks outperformed Bitcoin by 70% year-to-date in 2026, according to Bitcoin.com. The total sector market cap stood at $77.6 billion.
| Company | YTD Stock Performance | |---------|----------------------| | TeraWulf | +85% | | Hut 8 | +67% | | Riot Platforms | +46% |
The disconnect reflects the market re-rating miners as AI infrastructure plays rather than Bitcoin-denominated revenue businesses. Investors are pricing in forward AI revenue from signed contracts ($70B+ aggregate), discounting current Bitcoin mining losses as a temporary cost of transition.
CleanSpark closed up $0.10 on the day it reported a $378.3 million Q1 loss. MARA traded up 6.6% the day it disclosed a $1.3 billion loss and $1.5 billion BTC sell-off. Both stocks remain well below their 52-week highs, according to Blockhead.
The CoinShares Mining ETF (WGMI) serves as a sector proxy. Its performance tracks the revaluation thesis: miners with signed hyperscaler contracts trade at premium multiples to power capacity, not to hashrate.
The economic migration of hashrate from Bitcoin to AI workloads raises structural questions about network security.
Hashrate concentration: The United States, China, and Russia collectively control approximately 68% of global hashrate, according to the Cambridge Centre for Alternative Finance. As U.S.-listed miners redirect power capacity to AI, the share of remaining hashrate in less transparent jurisdictions may increase.
Difficulty spiral risk: Six negative difficulty adjustments in 2026 have reduced the cost of a theoretical 51% attack. While the current hashrate of approximately 1.03 ZH/s (as of May 2026) remains orders of magnitude above any practical attack threshold, the directional trend — declining hashrate during a period of rising total crypto market capitalization — is historically unusual.
Energy reallocation: Natural gas now powers 38.2% of Bitcoin mining, according to the 2025 CCAF report, while sustainable sources account for 52.4% (including 9.8% nuclear and 42.6% renewables). As miners redirect gas-powered facilities to AI, the remaining mining fleet skews toward renewable and stranded energy sources — potentially improving mining's environmental profile while reducing total security budget.
CoinShares forecasts hashrate growth to 1.8 ZH/s by end of 2026, conditional on Bitcoin recovering toward $100,000. The current trajectory suggests that projection requires a significant price reversal.
The Bitcoin mining industry is undergoing its most fundamental transformation since the 2021 China ban, but this time the migration is voluntary and economically rational. Miners are not fleeing regulation; they are following capital. The $70 billion in signed AI contracts dwarfs the approximately $8 billion in annual Bitcoin mining revenue the sector generated in 2025.
The economic logic is straightforward: a megawatt of power capacity generates more predictable, higher-margin revenue when leased to a hyperscaler for GPU workloads than when deployed against Bitcoin's difficulty-adjusted mining algorithm at current prices. That calculus reverses if Bitcoin reaches $100,000+, but miners cannot wait for a price recovery while servicing debt at current levels.
The open question is whether Bitcoin's security model can tolerate a sustained period of hashrate stagnation or decline. The network has survived worse — the 2021 China ban removed approximately 50% of hashrate overnight, and the protocol's difficulty adjustment mechanism restored equilibrium within months. The current 11% decline is modest by comparison. But the 2021 recovery was driven by miners relocating to the U.S. and plugging back in. This time, the power infrastructure itself is being repurposed for non-mining uses, making a rapid hashrate recovery conditional on new capital expenditure rather than simple reconnection.
For the foreseeable future, the public mining sector is a hybrid: part Bitcoin miner, part AI infrastructure REIT. The market has already priced in that transition. Whether the Bitcoin network can sustain its security guarantees during this period of divided attention is a question the mining companies themselves are no longer financially incentivized to answer.