Bitcoin mining is experiencing its deepest profitability crisis since China's 2021 crackdown. Hashprice — the industry's core revenue metric — has collapsed to approximately $35 per PH/s per day, an all-time low, while network difficulty surged 15% on February 20 in its largest single adjustment ...
"All miners should now be actively transitioning from BTC to HPC if at all possible." — Chris Brendler, Senior Analyst, Rosenblatt Securities
Bitcoin mining is experiencing its deepest profitability crisis since China's 2021 crackdown. Hashprice — the industry's core revenue metric — has collapsed to approximately $35 per PH/s per day, an all-time low, while network difficulty surged 15% on February 20 in its largest single adjustment since 2021. The result: revenue per terahash has fallen below $0.03, rendering the vast majority of mining operations unprofitable.
But this is not a cyclical downturn that will self-correct with the next price rally. A structural transformation is underway. The same infrastructure that once secured Bitcoin's network — cheap power, land, cooling systems, and grid interconnects — has become enormously valuable to a wealthier, less flexible buyer: the artificial intelligence industry. Hut 8 has signed a $7 billion, 15-year AI data center lease backstopped by Google. Cipher Mining has locked in a $5.5 billion, 15-year contract with AWS. TeraWulf has secured $6.7 billion in AI-related commitments. For miners who successfully pivot, mining revenue is projected to plummet from 85% of total revenue in early 2025 to less than 20% by end of 2026. The question facing the industry — and every Bitcoin holder — is what happens to network security when the infrastructure that protects $1.3 trillion in value is permanently rerouted to serve Anthropic and Google.
The scale of the mining industry's economic deterioration is best understood through its key performance indicators:
The economics are brutal at the rig level. According to CryptoSlate analysis, electricity costs now represent 52% of revenue even for the most efficient hardware available — Bitmain's Antminer S21 XP Hydro units. For mid-generation rigs like the S19 XP and Avalon A1466i, electricity consumes 92% to 100% of revenue, leaving zero margin for facility costs, labor, or capital depreciation. Older hardware — the Avalon A1366 and S19 Pro — is operating at electricity-cost-to-revenue ratios of 109% to 162%. These machines are literally burning cash.
Three forces have converged to create this crisis:
1. Post-Halving Revenue Compression. The April 2024 halving cut the block subsidy from 6.25 BTC to 3.125 BTC. Miners now earn half the Bitcoin per block that they did two years ago. While the market anticipated a price rally to compensate — and got one, briefly — Bitcoin's subsequent decline from $126,500 to $68,000 has erased the offset entirely.
2. The Winter Storm Shock and Difficulty Whiplash. A severe winter storm across the United States in early February forced several major mining operations to shut down, causing hashrate to plummet from 1.0 ZH/s to approximately 826 EH/s — a 12% decline and the steepest drawdown since China's 2021 ban. When operations restarted and hashrate recovered to 1.0 ZH/s, difficulty adjusted upward by 15%, compounding the profitability squeeze on already-marginal operators.
3. AI Competition for Power. Perhaps the most significant and irreversible factor: AI data centers are now aggressively outbidding Bitcoin miners for access to cheap electricity. These facilities offer energy providers long-term, firm-power contracts with higher margins and political support that miners cannot match. BlackRock's 2026 Global Outlook estimates that AI data centers could consume as much as 24% of total U.S. electricity by 2030, part of a $5 trillion to $8 trillion capital spending cycle that is fundamentally repricing the energy market.
The mining industry's response to this crisis has been swift and dramatic. Rather than fight for increasingly expensive electrons, the largest operators are converting their most valuable asset — power infrastructure — into AI computing capacity.
The deals are staggering in scale:
| Company | AI Partner | Deal Value | Capacity | Term | |---------|-----------|------------|----------|------| | Hut 8 | Google/Fluidstack (for Anthropic) | $7.0B | 245+ MW | 15 years | | TeraWulf | Google Cloud | $6.7B | Multiple sites | Multi-year | | Cipher Mining | AWS | $5.5B | 300 MW | 15 years | | Bitfarms | — | Full pivot | All BTC ops winding down | — |
Hut 8 CEO Asher Genoot called the Google-backed Fluidstack deal "the first domino to fall," signaling that the conversion of mining infrastructure to AI is just beginning. The company's River Bend campus in Louisiana will deliver at least 245 MW of AI data center capacity, with expansion options up to 2,295 MW that could lift total contract value to $17.7 billion.
The financial logic is overwhelming. AI data center operations generate 80% to 90% operating margins. For companies that have secured contracts, mining revenue is projected to collapse from 85% of total revenue in early 2025 to under 20% by end of 2026. Rosenblatt has given Overweight ratings to Cipher Mining and TeraWulf for their successful transitions, while issuing an Underweight on MARA Holdings for its continued full exposure to Bitcoin price volatility.
Bitfarms has gone furthest, announcing it will completely wind down Bitcoin mining operations and pivot entirely to AI — a sentence that would have been incomprehensible to the industry just 18 months ago.
Here is where the economic-value analysis becomes existential for Bitcoin.
Jeff Feng, co-founder of Sei Labs, described the current period as "the biggest bitcoin miner capitulation since 2021." CryptoQuant's profit-and-loss sustainability index has slumped to 21 — its lowest reading since late 2024. The network's total hashrate has declined approximately 12% since November 2025, leaving it at its weakest level since September 2025.
But the critical distinction between this cycle and previous ones is permanence. In past downturns, marginal miners shut down temporarily, difficulty adjusted downward, and survivors became more profitable. The cycle self-corrected. This time, the infrastructure itself is being permanently reallocated.
Once a 245 MW site is re-racked with GPUs under a 15-year AI lease, that power capacity is — in practice — permanently unavailable for future hashrate expansion. It is not mothballed mining capacity waiting for higher Bitcoin prices. It is gone.
This creates a paradox at the heart of Bitcoin's security model. The network relies on miners investing in hashrate to secure $1.3 trillion in value. But miners are economically rational actors, and AI infrastructure currently offers dramatically superior returns. If sufficient mining capacity migrates permanently, Bitcoin's security budget — already dependent on $18.1 billion annually in inflationary block subsidies — faces a structural deficit that no price rally alone can solve.
The foundational economic reality remains: Bitcoin requires approximately $54–72 billion annually in combined subsidies and infrastructure costs to secure a network that generates only ~$115 million in organic transaction fee revenue. The AI pivot is not creating this sustainability gap, but it is accelerating the timeline at which it becomes critical.
BlackRock's 2026 Global Outlook explicitly addressed this tension. The investment giant warned that AI-driven data centers present a fundamentally different power consumer than Bitcoin mining:
The implication is clear: in any competition for the same megawatt, AI wins. And with $5–8 trillion in AI infrastructure capital expenditure projected through 2030, this competition is only intensifying.
The mining industry is bifurcating into two categories:
Winners — The Infrastructure Converters: Companies like Hut 8, TeraWulf, and Cipher Mining that possess power assets valuable enough to attract tier-one AI customers. These firms are being re-rated by Wall Street not as miners but as AI infrastructure plays, with projected revenue growth of 66% to 132% in 2026.
Losers — The Pure-Play Miners: Companies that remain fully exposed to Bitcoin's price and mining economics. MARA Holdings, the largest public miner by hashrate, is down 13% year-to-date and carries an Underweight rating from Rosenblatt. Bitmine Immersion Technologies has declined 29% in 2026. For pure-play miners without convertible infrastructure, the path forward narrows with each difficulty adjustment.
The endgame is consolidation. The surviving pure-play miners will be those with the absolute lowest energy costs — likely in jurisdictions with stranded hydro, geothermal, or flared gas resources. Everyone else will either convert to AI, merge, or shut down.
The Bitcoin mining industry is undergoing the most profound structural transformation in its history. This is not a bear market stress test — it is an identity crisis. The physical infrastructure that once existed solely to secure the world's largest cryptocurrency is being permanently repurposed to train AI models and run inference workloads, because the economics demand it.
For investors, the signal is clear: the "Bitcoin miner" category is dissolving. What emerges in its place are AI infrastructure companies that happen to mine some Bitcoin on the side. The market is already pricing this in — companies with AI contracts trade at dramatic premiums to pure-play miners.
For Bitcoin itself, the implications are more profound and less resolved. The network's security model was designed for a world where miners had no alternative use for their infrastructure. That world no longer exists. Whether Bitcoin's fee market can mature fast enough to sustain security spending without inflationary subsidies — now competing against $7 billion Google-backed leases for the same megawatts — is the most important economic question in crypto today.