Bitcoin mining is experiencing its most severe stress event since China's 2021 ban. On February 9, the network executed an 11.16% downward difficulty adjustment — the largest in nearly five years — as hashrate plunged 20% from its October 2025 peak of 1.1 ZH/s. The Hash Ribbon indicator, which tr...
"There is no CEO of Bitcoin. There will be no bailout." — Matthew Sigel, Head of Digital Asset Research, VanEck
Bitcoin mining is experiencing its most severe stress event since China's 2021 ban. On February 9, the network executed an 11.16% downward difficulty adjustment — the largest in nearly five years — as hashrate plunged 20% from its October 2025 peak of 1.1 ZH/s. The Hash Ribbon indicator, which tracks miner capitulation by comparing short- and long-term hashrate moving averages, has been flashing distress for almost 60 consecutive days. With Bitcoin trading near $68,800 — down 45% from its $126,198 all-time high — and hashprice collapsing to $33 per petahash per day (from $70 at peak), an entire tier of the mining industry is being forced to choose: pivot, sell, or shut down.
This is not merely a cyclical squeeze. The 2024 halving cut block rewards to 3.125 BTC, permanently halving miner revenue per block. At the same time, a parallel economic opportunity — AI compute infrastructure — is pulling the most sophisticated operators away from Bitcoin entirely. Companies like IREN, Hut 8, and Core Scientific have signed multi-billion-dollar AI hosting deals that dwarf their mining economics. What's unfolding is a structural bifurcation of the mining industry: those who can convert megawatts into AI revenue will survive; those who cannot will be absorbed or liquidated. The implications for Bitcoin's security model, supply dynamics, and price are profound.
The data paints an unambiguous picture of industry-wide distress:
| Metric | Peak (Oct 2025) | Current (Feb 2026) | Change | |--------|-----------------|---------------------|--------| | Network Hashrate | 1,082 EH/s | ~860 EH/s | -20% | | Mining Difficulty | 141.6T | 125.86T | -11.16% | | Hashprice ($/PH/s/day) | ~$70 | ~$33 | -53% | | Bitcoin Price | $126,198 | ~$68,800 | -45% | | Block Reward | 3.125 BTC | 3.125 BTC | Fixed post-halving |
The difficulty drop from 141.6 trillion to 125.86 trillion represents miners physically unplugging machines. At $33 per petahash per day, CleanSpark's total cash-based mining cost of $30/PH/s and IREN's $26/PH/s leave razor-thin margins even for the most efficient public operators. Miners running older-generation ASICs or paying electricity rates above $0.06/kWh are operating at a loss.
Winter Storm Fern in late January compounded the crisis, forcing up to 40% of the network's hashrate offline as Texas grid operators issued curtailment orders. Some companies reported daily Bitcoin production falling by more than 60% during the storm's peak. But the weather was the accelerant, not the cause. The structural problem is economic: mining one Bitcoin now requires approximately 854,400 kWh of electricity — equivalent to 81 years of average U.S. residential consumption.
The April 2024 halving cut Bitcoin's block subsidy from 6.25 to 3.125 BTC — a permanent 50% revenue reduction per block. At the time, Bitcoin traded above $60,000, providing a temporary cushion. But the math was always unforgiving: miners needed price appreciation to offset the halving's revenue impact.
Instead, they got a 45% drawdown. According to MEXC Research, production costs for many industrial miners have risen to approximately $87,000 per Bitcoin, while the spot price sits near $68,800. This gap — a negative margin of roughly $18,000 per coin — cannot be sustained. The ROI period for new mining hardware has stretched to approximately 1,000 days, meaning most new rigs won't pay for themselves before the next halving in 2028.
VanEck's Matthew Sigel identified five factors driving the February selloff: collapsing leverage (futures open interest dropped from $61 billion to $49 billion in one week), AI sector contagion, forced miner selling, quantum computing narrative risk, and the typical four-year cycle psychology. The February 5 crash registered a -6.05σ move on the rate-of-change Z-score — placing it among the fastest single-day price collapses in Bitcoin's history.
JPMorgan's mining research team, led by Reginald Smith and Charles Pearce, responded by trimming price targets on MARA Holdings (from $20 to $13) and Riot Platforms (from $19 to $17), while upgrading Cipher Mining and CleanSpark — both companies with credible AI infrastructure pivots.
The most consequential development in Bitcoin mining is not about Bitcoin at all. It is the mass migration of mining infrastructure toward artificial intelligence workloads — and the numbers explain why.
The Deals That Changed Everything:
JPMorgan expects miners to announce roughly 1.7 gigawatts of additional AI-related capacity by late 2026 — equal to about 35% of their total approved power footprint. The logic is straightforward: AI hyperscalers offer longer-term, more predictable revenue contracts than Bitcoin mining. NVIDIA Blackwell GPUs are sold out through mid-2026, with a backlog of 3.6 million units. Companies that secured early allocations hold a "hardware moat" that protects their revenue projections.
This is economic natural selection at work. Mining companies are not technology companies — they are energy arbitrage companies. Their core competency is securing cheap power at scale. AI simply pays more for that power than Bitcoin does.
Public mining companies hold significant Bitcoin treasuries, creating a latent source of forced selling:
| Company | BTC Holdings (Dec 2025) | Deployed Hashrate | |---------|------------------------|-------------------| | MARA Holdings | 44,893 BTC | 53.2 EH/s | | Riot Platforms | 17,722 BTC | 31.5 EH/s | | CleanSpark | 9,952 BTC | 35.5 EH/s |
Total miner reserves have fallen to approximately 1.806 million BTC. When miners operate at negative margins, they face a cascading sequence: sell Bitcoin holdings to cover operating costs, then reduce hashrate, then — if price doesn't recover — face creditor-forced liquidation of remaining holdings.
This is already happening. Cango Inc. liquidated 4,451 BTC ($305 million) to deleverage and fund its transition to GPU-based AI computing. Russia's BitRiver, one of the largest mining operations outside the U.S., faces insolvency proceedings as its CEO was placed under house arrest. CleanSpark reported a $378.7 million net loss for Q1 2026 due to Bitcoin devaluation and is actively pivoting capital expenditure toward AI.
The risk of a "doom loop" is real: forced miner selling pushes price lower, which pushes more miners below breakeven, which triggers more selling. In February alone, crypto markets experienced $2.65 billion in liquidations over a single 24-hour period, with 586,053 traders liquidated.
Bitcoin's difficulty adjustment algorithm — often called its most underappreciated feature — acts as a built-in circuit breaker. When unprofitable miners shut down, difficulty drops, which immediately improves margins for surviving miners. The February 11.16% adjustment represents exactly this mechanism in action.
Historically, every major capitulation event has preceded significant price rallies:
VanEck's research team explicitly characterized the current capitulation as "auspicious for the $2 trillion digital asset," noting that the pattern of weaker miners exiting while stronger ones absorb capacity has historically set the stage for price expansion.
The contrarian case is structural: when hashrate drops and difficulty adjusts downward, surviving miners become more profitable at the same price. The network doesn't need all miners — it needs enough. At 860 EH/s, Bitcoin remains the most computationally secure network in history by orders of magnitude.
The mining industry is splitting into three tiers:
Tier 1: AI-Pivoted Infrastructure Companies IREN, Core Scientific, Hut 8, TeraWulf. These companies have secured multi-year, multi-billion-dollar AI hosting contracts that provide predictable revenue regardless of Bitcoin price. Mining becomes a secondary revenue stream — or is abandoned entirely.
Tier 2: Efficient Pure-Play Miners CleanSpark, Cipher Mining. Operating at or near the efficiency frontier ($26-30/PH/s cash costs), these companies survive by maintaining the lowest possible cost structure while exploring AI optionality. JPMorgan upgraded both.
Tier 3: Undifferentiated Miners Companies without AI deals, older hardware, or high energy costs. These are the miners being capitulated. Their exit is what the Hash Ribbon is measuring.
The separation is permanent. The AI infrastructure buildout requires 18-24 months of lead time for site preparation, grid interconnection, and GPU procurement. Companies that missed the 2025 window for Blackwell allocations and hyperscaler negotiations cannot catch up. MARA Holdings, despite being the largest public miner by hashrate, was labeled a "laggard" in the HPC pivot by industry analysts, having pursued a "strategic pause" that allowed peers to capture early allocations.
The Bitcoin mining industry is undergoing a structural transformation that will permanently alter its composition. What began as a cyclical squeeze — halving-reduced revenue meeting a 45% price drawdown — has become an existential fork. The AI compute opportunity is not a distraction; it is a repricing of the same physical infrastructure (power, cooling, real estate) that miners already control. Companies that convert megawatts from $33/PH/s hashprice into multi-billion-dollar AI contracts are making a rational economic decision.
For Bitcoin's network, the immediate risk is real but manageable. The protocol's difficulty adjustment ensures that mining remains viable for some operators at any price level above zero. The concentration of hashrate among larger, better-capitalized miners may reduce decentralization, but it also makes the remaining network more economically resilient.
For investors, the Hash Ribbon signal carries a strong historical track record as a contrarian indicator. Every previous capitulation of this magnitude has preceded significant price appreciation — not because capitulation is bullish, but because it represents the final phase of forced selling before supply exhaustion.
The miners who survive this reckoning will be fundamentally different companies than they were a year ago. Many will derive more revenue from AI than from Bitcoin. The question is no longer whether mining is profitable. It is whether mining is the best use of the energy these companies control. For a growing number of them, the answer is no.