Bitcoin mining economics have deteriorated to their worst levels since the April 2024 halving. Hash price — the revenue a miner earns per unit of computing power — fell to $27.89 per petahash per day in mid-2026, down 66% from the October 2025 peak of approximately $82/PH/day. Network hashrate ha...
"Right now [investing more in Bitcoin mining] doesn't make sense. So we want to be redirecting every dollar possible toward AI capex." — Gary Vecchiarelli, President and CFO, CleanSpark
Bitcoin mining economics have deteriorated to their worst levels since the April 2024 halving. Hash price — the revenue a miner earns per unit of computing power — fell to $27.89 per petahash per day in mid-2026, down 66% from the October 2025 peak of approximately $82/PH/day. Network hashrate has declined 23% from its 1.1 ZH/s peak to approximately 844-893 EH/s as of early July 2026, the largest peak-to-trough decline since China's 2021 mining ban.
The industry response has been binary: sell bitcoin to fund an exit from mining, or die. Publicly traded miners liquidated more than 32,000 BTC in Q1 2026 alone — a single-quarter record exceeding combined sales for all of 2025. The capital is flowing into AI and high-performance computing (HPC) infrastructure, where more than $70 billion in cumulative contracts have been announced across the public mining sector. Core Scientific now earns 67% of its revenue from AI colocation. IREN raised $3 billion to install 140,000 GPUs. The Bitcoin mining industry, as a standalone sector, is ceasing to exist in its prior form.
The April 2024 halving reduced Bitcoin's block subsidy from 6.25 BTC to 3.125 BTC. Operating costs did not halve with them. According to the CoinShares Q1 2026 Mining Report, the weighted average cash cost to produce one bitcoin among publicly listed miners rose to approximately $79,995 in Q4 2025 — a period when BTC traded between $86,000 and $124,500.
The math worsened through H1 2026. Key metrics as of July 2026:
| Metric | Value | Change from Peak | |--------|-------|-----------------| | Hash price | $27.89–$33/PH/day | -66% from Oct 2025 | | Network hashrate | ~844–893 EH/s | -23% from 1.1 ZH/s peak | | Mining difficulty | 124.93T (Jun 13 low) | -20% from ATH | | Avg. cost per BTC (industrial) | $32,000–$55,000 (power only) | N/A | | BTC price (Jul 10) | ~$62,150 | -50% from $124,500 ATH |
At current hash prices, miners running mid-generation hardware (28 J/TH weighted average efficiency) require sub-$0.05/kWh electricity to remain cash-profitable. Only latest-generation fleets operating below 15 J/TH retain meaningful margins at typical industrial rates ($0.06–$0.08/kWh). CoinShares estimates 15–20% of legacy rigs are now operating at a loss.
JPMorgan confirmed in a research note that BTC was trading approximately 19% below the average production cost of listed miners, characterizing the environment as "structurally unprofitable for marginal producers."
The network's self-correcting difficulty mechanism has confirmed the exodus. Bitcoin difficulty dropped 10.09% at block 953,568 on June 13, 2026, reaching 124.93T — the lowest level in 11 months. This followed a pattern of negative adjustments: seven of the last eight difficulty retargets in early 2026 were downward.
The next difficulty adjustment is estimated for July 11, 2026, with average block times running 35 seconds slower than the 10-minute target — indicating continued hashrate departure.
This represents the most sustained period of miner capitulation since the 2021 Chinese mining ban, when approximately 50% of global hashrate went offline within weeks. The current decline is slower but potentially more structural: miners are not relocating, they are converting their facilities to different workloads.
Publicly traded Bitcoin miners sold more than 32,000 BTC in Q1 2026, per Galaxy Research. This exceeded their combined sales for all of 2025. The sales were not panic liquidations — they were strategic divestitures to fund AI infrastructure buildouts.
Notable liquidation events:
Galaxy Research confirmed on June 21, 2026 that the industry had entered a formal capitulation phase — defined as operators being forced off the network by losses rather than choosing to exit strategically. However, the distinction between forced exit and strategic pivot has blurred: many miners are simultaneously unprofitable at mining and actively choosing to reallocate capital toward higher-margin AI workloads.
The conversion of Bitcoin mining infrastructure to AI data centers has crossed from strategy slide decks to construction sites. According to industry data compiled by insights4vc, more than $70 billion in cumulative AI and high-performance computing contracts have been signed by former or current Bitcoin mining companies.
The economic logic is straightforward: AI data centers use the same power and cooling infrastructure as mining operations but generate higher revenue per kilowatt-hour when electricity prices exceed $0.05/kWh. TeraWulf CEO Paul Prager framed the shift explicitly: "We are fundamentally a power company that builds digital infrastructure, not the other way around."
Industry projections suggest AI and HPC could account for 70% of revenue for transformed miners by end of 2026. The publicly traded mining sector has outperformed BTC year-to-date, with gains ranging from 5% to 85% — driven entirely by AI contract valuations, not mining economics.
Key AI infrastructure deals announced by mining companies:
| Company | Deal Value | Partner | Capacity | |---------|-----------|---------|----------| | Core Scientific | $10.2B (12-year) | CoreWeave | 590 MW | | IREN | $3B convertible notes | Microsoft | 4.5 GW pipeline | | CoreWeave (miner-hosted) | $11.9B (5-year) | OpenAI | N/A |
The clearest example of complete transformation. Q1 2026 results:
Core Scientific is converting all 11 facilities across seven states to support AI workloads. The company's mining revenue now represents a legacy runoff portfolio.
Raised $3 billion in convertible notes in May 2026 to fund AI transition. Targets:
However, IREN's Q3 FY2026 revenue of $144.8 million missed analyst expectations of $219.87 million, indicating the transition timeline carries execution risk.
Reported Q1 2026 revenue of $174.6 million, down 18.3% year-over-year, with a net loss of $1.3 billion. Marathon CEO Fred Thiel stated the objective is to "direct computational power toward its most productive use."
Quarterly revenues fell 24.9% to $136.4 million. Net loss widened to $378.3 million from $138.8 million in the year-ago quarter. CFO Gary Vecchiarelli's earnings call statement — that every available dollar should go toward AI capex — represents the starkest admission from a company that previously positioned itself as a pure-play miner.
The 23% hashrate decline raises questions about Bitcoin's security model, though context is required. At ~893 EH/s, the network remains orders of magnitude more secure than at any point before 2024. The difficulty adjustment mechanism is functioning as designed — reducing difficulty to restore profitability for surviving miners and maintain ~10-minute block times.
The more structural concern is concentration. An estimated 85% of global hashrate is now controlled by approximately 12 publicly traded entities or sovereign wealth funds, according to CoinShares. This concentration, combined with the AI pivot, creates a scenario where Bitcoin's security budget competes directly with AI compute demand for the same physical infrastructure.
If AI workloads consistently offer higher returns per kWh than Bitcoin mining at prevailing hash prices, rational operators will continue reallocating capacity away from the network. The equilibrium depends on BTC price recovery sufficient to restore mining margins — or transaction fee revenue growth from increased on-chain activity.
The Bitcoin mining industry is undergoing a structural transformation that transcends a normal halving cycle. Previous halvings (2016, 2020) triggered temporary margin compression followed by price recovery that restored profitability. The 2024 halving coincided with the emergence of a competing demand source for the same infrastructure — AI compute — that offers superior and more predictable unit economics.
The result is not a temporary shakeout but a permanent reallocation of capital and infrastructure. Companies that once competed to accumulate hashrate now compete to sign GPU hosting contracts. The 32,000 BTC liquidated in Q1 2026 was not distressed selling — it was funding for data center conversions that will never revert to mining.
For Bitcoin's network, the implications are manageable in the near term: difficulty adjustments ensure block production continues. The longer-term question is whether the security budget — currently subsidized by the 3.125 BTC block reward — can sustain itself as that subsidy halves again in 2028, particularly if the physical infrastructure that once secured the network has been permanently repurposed for artificial intelligence.