Public Bitcoin miners sold 32,000 BTC in Q1 2026 — more than in all four quarters of 2025 combined. The proceeds are financing a sector-wide pivot to artificial intelligence infrastructure. Over $70 billion in AI and high-performance computing (HPC) contracts have been signed by listed mining fir...
"We are no longer a Bitcoin company." — Ben Gagnon, CEO, Bitfarms
Public Bitcoin miners sold 32,000 BTC in Q1 2026 — more than in all four quarters of 2025 combined. The proceeds are financing a sector-wide pivot to artificial intelligence infrastructure. Over $70 billion in AI and high-performance computing (HPC) contracts have been signed by listed mining firms, and according to Bloomberg, AI revenue is on track to surpass Bitcoin mining revenue for leading operators by year-end, with projections reaching 70% of total revenue by December 2026, up from approximately 30% in Q1.
The economics are straightforward. CoinShares' Q1 2026 mining report places the average production cost per BTC among listed miners at approximately $79,995, while Bitcoin traded between $68,000 and $75,000 during the same period. Hashprice — the daily revenue per unit of computing power — collapsed to $28-30 per PH/s/day in Q1 2026, a new historical low since the April 2024 halving. Between 15% and 20% of the global mining fleet is now operating at a loss. The response has been uniform: convert power infrastructure to GPU compute, sign hyperscaler leases, and sell bitcoin to fund the transition.
Mining stocks have outperformed BTC by roughly 70% year-to-date, reflecting market consensus that the AI pivot creates more value than block rewards.
The April 2024 halving cut block rewards from 6.25 BTC to 3.125 BTC. Post-halving compression was expected. What was not anticipated was the sustained decline in hashprice throughout 2025 and into 2026.
According to CoinShares' Q1 2026 Bitcoin Mining Report, hashprice peaked at approximately $63/PH/s/day in July 2025. It declined steadily through Q4 2025, dropping to $35-37/PH/s/day by November — a five-year low at the time. A brief rebound to $38-40 in late December was short-lived. By Q1 2026, hashprice collapsed further to $28-30/PH/s/day, a new post-halving record low.
At this hashprice, any miner running hardware below an Antminer S19 XP with electricity costs at or above $0.06/kWh is losing money. CoinShares estimates this covers 15% to 20% of the global mining fleet. Marathon Digital reported electricity costs of $64,703 per BTC, reflecting heavy reliance on third-party hosting. The weighted average cash cost across listed miners reached approximately $79,995 per BTC in Q4 2025 — above Bitcoin's trading range of $68,000-$75,000 during the same period.
The math is clear: at current prices and difficulty levels, pure-play Bitcoin mining is a negative-margin business for most publicly traded operators.
The pivot is not speculative. Over $70 billion in cumulative AI and HPC contracts have been announced across the public mining sector. The deals are anchored by hyperscalers — Microsoft, Google (via its cloud partners), CoreWeave, and Anthropic — seeking the exact infrastructure miners already possess: high-density power, cooling systems, and fiber connectivity in locations where grid capacity is available.
The largest contracts by value:
| Company | Counterparty | Contract Value | Duration | Capacity | |---------|-------------|----------------|----------|----------| | Core Scientific | CoreWeave | $10.2B | 12 years | 590 MW across 6 sites | | TeraWulf | Fluidstack/Core42 (Google-backed) | $12.8B | Multi-year | Undisclosed | | Hut 8 | Anthropic/Fluidstack | $7B | 15 years | River Bend campus | | IREN | Microsoft | $1.94B annualized | 5 years | 200 MW liquid-cooled GPU |
IREN's Microsoft deal carries an 85% project-level EBITDA margin — substantially above Bitcoin mining margins, which have turned negative for many operators.
Core Scientific's AI colocation revenue already accounts for 39% of total revenue. IREN's HPC revenue is projected to surge from 3% of total revenue in 2024 to 71% by year-end 2026. TeraWulf is expected to reach 70% AI/HPC revenue from near-zero in 2024. The speed of this transition is without precedent in the mining sector.
Public miners collectively sold approximately 32,000 BTC in Q1 2026, according to Cointelegraph and KuCoin Research. This exceeds total miner BTC sales across all four quarters of 2025 and surpasses the previous record of roughly 20,000 BTC sold in Q2 2022 during the Terra-Luna collapse.
The sales are funding AI infrastructure buildouts. Notable transactions:
The shift from "HODL" treasury strategy to systematic liquidation marks a structural change in how mining companies view their relationship with Bitcoin. The asset has become working capital for an AI buildout, not a long-term store of value on corporate balance sheets.
Core Scientific (CORZ): The furthest along in the transition. AI colocation at 39% of revenue. CoreWeave partnership worth $10-12 billion across 590 MW at six sites, including a $1.2 billion expansion in Denton, Texas. Stock up over 40% YTD.
TeraWulf (WULF): Leads public miners with a 73.58% YTD stock gain. $12.8 billion in contracted HPC revenue through Google-backed Fluidstack and Core42. AI/HPC at 27% of revenue and climbing.
Hut 8 (HUT): Trading at $77.06, the highest share price among the top ten listed miners. Anchored a $7 billion, 15-year lease at its River Bend campus with Anthropic and Fluidstack. Building an 8.5 GW development pipeline across multiple construction stages.
IREN (IREN): Stock advanced to $52.02 on April 23 after positive analyst coverage. The Microsoft five-year deal projects $1.94 billion in annualized revenue at 85% project-level EBITDA margin. Up over 750% in the past year. Currently building 200 MW of liquid-cooled GPU capacity. Deploying NVIDIA GB200 NVL72 systems.
Bitfarms: CEO Ben Gagnon declared the company is "no longer a Bitcoin company." Doubling down on AI infrastructure in the U.S.
CleanSpark: Announced a "business evolution from pure-play Bitcoin miner to include AI compute." Acquired 447 acres in Brazoria County, Texas, for a large-scale data center with plans for 600 MW of capacity.
Bitdeer: Liquidated entire Bitcoin treasury. Deploying NVIDIA GB200 NVL72 systems in Malaysia and converting sites in the U.S. and Europe from crypto mining to AI data centers.
Despite the economic squeeze on miners, Bitcoin's network hashrate remains near record levels. The network held at approximately 994.76 EH/s across the April 3-17 observation window, just below the 1 ZH/s threshold first breached in early 2026.
Mining difficulty eased on April 17, 2026, declining 2.43% to 135.59 trillion — the fifth downward revision of the year. Of eight total difficulty adjustments in 2026, five have been reductions and three increases. The average block interval of 9 minutes 35 seconds suggests the next adjustment on May 2 could move upward.
Hashprice has recovered modestly, climbing 13.65% between March 18 and April 18, providing some relief. However, the structural trend is clear: miners are not returning displaced hashrate. The capacity being converted to AI compute is permanent infrastructure reallocation.
The implications for Bitcoin's security model are not yet acute — hashrate remains near all-time highs — but the direction is notable. If the economic incentive to mine continues to deteriorate relative to AI compute, the network could eventually face hashrate plateaus or declines that would have been unthinkable two years ago.
The market has drawn a clear line. Mining stocks with credible AI pivots have outperformed BTC by approximately 70% in 2026. The top ten publicly listed miners posted YTD gains of 25-73%, while BTC itself sat roughly 12% in the red from January 1 through mid-April.
TeraWulf leads at 73.58% YTD. Core Scientific exceeds 40%. IREN is up over 750% on a trailing 12-month basis.
The valuation framework has shifted. Analysts at S&P Global, Benchmark, and CoinShares now assess mining companies primarily on their AI contract backlog, power pipeline, and GPU deployment timelines — not on hashrate, BTC production, or HODL strategy. The mining thesis, as an equity story, has been replaced by a data center thesis.
The pivot carries risks that are not reflected in current equity valuations:
Execution risk. Converting mining facilities to Tier 3+ data centers requires significant capital expenditure, specialized cooling, and uptime guarantees that differ from mining operations. Not all miners have the engineering capability.
Counterparty concentration. Several companies depend on a single hyperscaler counterparty (CoreWeave for Core Scientific, Microsoft for IREN). Contract termination or renegotiation could materially impact revenue projections.
AI compute demand cyclicality. Current AI infrastructure demand is driven by a buildout phase. If AI capital expenditure cycles downward — as some analysts have speculated — miners-turned-data-center-operators could face the same overcapacity they experienced in Bitcoin mining.
Bitcoin network security. The systematic diversion of power infrastructure from mining to AI reduces the long-term incentive to secure the Bitcoin network. This is not yet a crisis, but the trend merits monitoring.
Debt accumulation. Companies are funding the transition through convertible notes, equity raises, and BTC sales. Bitdeer alone raised $368.5 million in a single quarter. Rising leverage in a capital-intensive buildout introduces financial fragility.
The Bitcoin mining industry is undergoing a one-way structural transformation. The economic logic is unavoidable: AI compute delivers higher margins than block rewards, hyperscalers are willing to sign multi-billion-dollar, multi-year contracts for the exact infrastructure miners possess, and Bitcoin's post-halving economics have made pure-play mining a negative-margin business for most listed operators.
What began as opportunistic diversification in 2024 has become a full-sector identity change. Bitfarms' CEO calls it "no longer a Bitcoin company." Bitdeer has zero BTC on its balance sheet. Core Scientific generates 39% of revenue from AI and is accelerating. The question is no longer whether miners will pivot, but how quickly the remaining holdouts follow.
For Bitcoin's network, the near-term security implications are manageable — hashrate remains at historic highs. The longer-term question is whether a fee-based security model can sustain the network as the subsidy continues to halve and the infrastructure operators who once competed for block rewards redirect their capacity toward training runs and inference workloads.
The market has already priced in the answer: mining stocks trade on AI contract backlog, not hashrate. The era of the Bitcoin miner, as equity markets understood it, is over. The era of the former miner turned data center operator has begun.