Public Bitcoin miners sold a record 32,000 BTC in Q1 2026 — exceeding total net sales for all of 2025 — as hashprice collapsed to $28–30 per PH/s/day, a five-year low. The sell-off funded an accelerating pivot: over $70 billion in cumulative AI and high-performance computing (HPC) contracts have ...
"We are pivoting substantially all of our capital allocation toward AI infrastructure. Bitcoin mining alone cannot sustain the business at current hashprice levels." — Adam Sullivan, CEO, Core Scientific, Q4 2025 Earnings Call
Public Bitcoin miners sold a record 32,000 BTC in Q1 2026 — exceeding total net sales for all of 2025 — as hashprice collapsed to $28–30 per PH/s/day, a five-year low. The sell-off funded an accelerating pivot: over $70 billion in cumulative AI and high-performance computing (HPC) contracts have been signed across the listed mining sector, with projections that AI will account for 70% of miner revenue by year-end, up from approximately 30% at the start of the year.
The structural shift is visible in every metric. Bitcoin's network hashrate peaked at 1,160 EH/s in October 2025 and has since declined to approximately 940 EH/s — a 19% drop. Mining difficulty is projected to fall 14–16% at the next adjustment on April 19, 2026. The weighted average cash cost to mine one BTC among listed miners reached $79,995 in Q4 2025, according to CoinShares, while Bitcoin's spot price has fallen from $126,000 in October 2025 to roughly $77,000 in mid-April 2026 — compressing margins to the point where 15–20% of the global fleet is mining at a loss.
This report examines the economics driving the exodus, the security implications of declining hashrate, and whether the mining industry is undergoing a temporary rebalancing or a permanent transformation.
Hashprice — the USD revenue a miner earns per petahash per second per day — is the single most important metric for mining economics. Its trajectory since the April 2024 halving tells a stark story.
Timeline:
At $30/PH/s/day, any operation running hardware older than a Bitmain S19 XP with electricity costs at or above $0.06/kWh is losing money. This affects roughly 15–20% of the global mining fleet, according to CoinShares' Q1 2026 Bitcoin Mining Report.
The breakeven electricity cost for current-generation hardware (S21, M60 series) sits at approximately $0.045–0.055/kWh. Operators above this threshold are burning cash. The electricity-only breakeven BTC price for mid-tier operators is approximately $74,000 — above the current spot price of roughly $77,000, leaving minimal margin after overhead, depreciation, and debt service.
The fully loaded cost to produce one BTC — including depreciation and stock-based compensation — has reached approximately $137,800 per BTC for the average listed miner. At a spot price roughly 44% below this figure, the industry's aggregate economics are underwater on a GAAP basis.
Publicly listed miners collectively sold more than 32,000 BTC in Q1 2026, per Tekedia and CoinShares data. This exceeds the previous quarterly record of roughly 20,000 BTC set during Q2 2022, amid the Terra-Luna collapse.
Notable sales by company:
| Company | BTC Sold (Q1 2026) | Estimated Proceeds | Context | |---|---|---|---| | Core Scientific | ~1,900 BTC (Jan) | ~$175M | Plans to sell "substantially all" remaining holdings | | Riot Platforms | 3,778 BTC | ~$290M | Record quarterly sale | | Bitdeer | Reduced to 0 BTC | Undisclosed | Treasury fully liquidated by Feb 2026 | | MARA Holdings | Undisclosed | Undisclosed | Reported $1.71B net loss in Q4 2025 |
The motivations are uniform: fund AI infrastructure buildout, service convertible debt, and cover operating losses from mining. CleanSpark stands as a notable exception, retaining its 13,513 BTC treasury and exploring bitcoin-backed credit lines and covered-call strategies as non-dilutive financing alternatives.
The sell pressure from miners adds approximately $2.4 billion in supply to the market in a single quarter — a material overhang for a market already contending with broader risk-off sentiment.
The conversion of mining infrastructure to AI/HPC hosting is no longer a thesis. It is a deployed strategy with signed, long-term contracts.
Headline deals:
| Company | AI/HPC Partner | Capacity | Contract Value | Duration | |---|---|---|---|---| | Core Scientific | CoreWeave | ~590 MW | $10.2B | 12 years | | TeraWulf | Undisclosed | 522 MW | $12.8B | Long-term | | MARA Holdings | Starwood | ~1 GW (near-term) | Undisclosed | Multi-year | | IREN | Multiple | 200 MW (under construction) | Undisclosed | Multi-year | | Riot Platforms | AMD | Data center lease | Undisclosed | 10 years |
The economic logic is straightforward. AI hosting generates approximately three times the return of Bitcoin mining on a per-megawatt basis, according to industry analysis. AI customers pay steady rents with profit margins of 80–90%, compared to mining margins that swing with hashprice and BTC spot.
CoreWeave's expanded relationship with Core Scientific — culminating in a reported $9 billion all-stock acquisition offer — illustrates the endgame. The acquirer is not buying hash power. It is buying permitted megawatts, fiber connectivity, and cooling infrastructure. The Bitcoin mining operation is incidental.
Riot Platforms posted record annual revenue of $647.4 million for 2025 but saw adjusted EBITDA collapse to $12.96 million from $463.19 million the prior year. The AMD data center lease became operational in January 2026, providing a stabilizing revenue stream. The company faces growing pressure from investors to accelerate its AI transition.
CoinShares projects that listed miners could derive 70% of total revenue from AI by end of 2026, up from roughly 30% at the start of the year. For Core Scientific specifically, AI colocation revenue already accounts for 39% of total revenue and grew 268% year-over-year.
Bitcoin's network hashrate peaked at 1,160 EH/s in October 2025. By mid-April 2026, it has fallen to approximately 940 EH/s — a 19% decline from peak. The trajectory has been volatile:
Mining difficulty rose 3.87% at the most recent adjustment (block 943,488) but is projected to fall 14–16% at the next adjustment on approximately April 19, 2026, based on current block intervals averaging 11 minutes 39 seconds — well above the 10-minute target.
Security implications:
A lower hashrate directly reduces the economic cost of a 51% attack. During the January trough, the theoretical cost of a double-spend attack fell 30–40%. However, several structural factors mitigate systemic risk:
The structural question is whether AI revenue permanently diverts megawatts away from mining, leading to a lower long-run hashrate equilibrium. If AI generates more stable, higher-margin returns per megawatt, rational operators will allocate accordingly. Bitcoin's security budget — block subsidy plus transaction fees — must compete with AI hosting revenue for every megawatt of capacity.
At the current block subsidy of 3.125 BTC per block (~450 BTC/day), daily miner revenue from subsidies alone is approximately $34.6 million at $77,000 BTC. Transaction fees add another $1–3 million per day. Total annualized security spend: roughly $13–14 billion. Whether this is sufficient to incentivize 900+ EH/s of sustained hashrate while AI alternatives offer 3x returns per MW remains an open question.
The divergence in sovereign approaches to Bitcoin mining crystallized in April 2026.
Bhutan has sold approximately 70% of its peak holdings of ~13,000 BTC, reducing its balance to 3,954 BTC (~$280.6 million), according to on-chain data reported by CoinDesk. The kingdom appears to have ceased mining as of approximately November 2024, with no mining inflows exceeding $100,000 recorded since. Some $215.7 million in BTC has exited Bhutan-linked wallets in 2026 alone. At current difficulty levels and a 3.125 BTC block reward, Bhutan's hydropower may generate more revenue from electricity exports to India than from mining.
United States formalized a Strategic Bitcoin Reserve in March 2026, with holdings estimated at 325,000–328,000 BTC. Multiple U.S. senators introduced a reserve-backed mining bill.
Brazil reintroduced legislation for a national Bitcoin reserve (RESBit) targeting accumulation of up to 1 million BTC over five years.
At least 13 countries now operate Bitcoin mining at the central government level, including the UAE, Russia, Iran, Kazakhstan, Ethiopia, and Paraguay. The trend toward sovereign mining is accelerating even as private miners exit — raising questions about whether nation-state actors will absorb hashrate shed by commercial operators pivoting to AI.
The margin compression is driving the most aggressive consolidation wave in mining history. 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.
Consolidation dynamics:
The M&A pipeline favors acquirers with low-cost power access and existing AI contracts. Miners without either face accelerated attrition if BTC remains below $100,000.
The Bitcoin mining industry is undergoing a structural transformation, not a cyclical correction. The halving reduced the block subsidy. The BTC price decline compressed revenue. AI provided an alternative use for every permitted megawatt. These three forces converged to produce a quarter in which miners sold more BTC than during the Terra-Luna crash and signed more AI hosting contracts than in the prior two years combined.
The question is no longer whether miners will pivot to AI. They already have. The question is what happens to Bitcoin's security budget when a growing share of the world's purpose-built power infrastructure optimizes for GPU throughput instead of SHA-256 hashes. At 940 EH/s, the network remains extraordinarily secure by any historical standard. But the trend line — declining hashrate, rising pool concentration, and a block subsidy halving again in 2028 — suggests that Bitcoin's long-term security model will depend increasingly on transaction fee revenue that has yet to materialize at scale.
The difficulty adjustment will do what it always does: recalibrate. But it cannot create economic incentives that don't exist. If AI hosting continues to generate 3x the return per megawatt, the mining industry's identity crisis will deepen, and Bitcoin's security budget will compete for capital in a market that increasingly has better options.