Bitcoin mining just recorded its most violent two-week swing in half a decade. On February 9, 2026, network difficulty fell 11% — the steepest drop since China's mining ban in 2021 — after Winter Storm Fern knocked 455 EH/s offline and cratered hashrate to a seven-month low of 663 EH/s. Eleven da...
"Scaling frontier AI infrastructure is, at its core, a power challenge. Through our partnership with Anthropic and Fluidstack, we are aligning power, data center design, and compute deployment into an integrated platform capable of delivering at gigawatt scale." — Asher Genoot, CEO, Hut 8
Bitcoin mining just recorded its most violent two-week swing in half a decade. On February 9, 2026, network difficulty fell 11% — the steepest drop since China's mining ban in 2021 — after Winter Storm Fern knocked 455 EH/s offline and cratered hashrate to a seven-month low of 663 EH/s. Eleven days later, on February 20, difficulty snapped back by 15% to 144.4 trillion, the largest single upward adjustment since 2021, as miners flooded back online and pushed hashrate past the 1 ZH/s threshold once again.
The whipsaw obscures something more structural: the Bitcoin mining industry is splitting in two. On one side, a shrinking cohort of pure-play miners is being crushed between record difficulty, post-halving block reward compression, and a hashprice languishing at multi-year lows near $23.9 per PH/s per day. On the other, a new class of hybrid infrastructure operators — Hut 8, Core Scientific, TeraWulf, IREN — is pivoting gigawatts of power capacity toward AI and high-performance computing (HPC) contracts worth billions of dollars over the next decade. This comparative analysis maps the economic fault line between these two models and asks the question the market is now pricing: can Bitcoin mining survive as a standalone business?
The numbers tell a story of a network under extreme stress. Between late January and mid-February 2026, Bitcoin's SHA-256 hashrate experienced a 40% drawdown — from peaks near 1.13 ZH/s down to 663 EH/s over the January 25–26 weekend — before staging an equally dramatic recovery to approximately 1 ZH/s by February 19.
The difficulty algorithm, which adjusts every 2,016 blocks (roughly two weeks), responded accordingly:
| Date | Difficulty | Change | Trigger | |------|-----------|--------|---------| | Feb 9, 2026 | ~125.9 T | -11% | Storm Fern curtailment; miner capitulation | | Feb 20, 2026 | 144.4 T | +15% | Hashrate recovery to 1 ZH/s |
The -11% / +15% oscillation within a single month is unprecedented in Bitcoin's post-2021 history. Foundry USA, the largest U.S. mining pool, lost an estimated 60% of its capacity during the peak of disruption, with approximately 200 EH/s going offline simultaneously. The recovery, while impressive, masks a deeper question: the miners that came back online are not the same mix that went offline.
Hashprice — the daily revenue a miner earns per petahash per second of deployed capacity — is the single most important metric in mining economics. As of February 20, 2026, it sits at approximately $23.9 per PH/s per day, representing a roughly 35% decline from approximately $55 per PH/s per day in mid-2025.
JPMorgan's mining research team estimates the all-in Bitcoin production cost at $77,000 per BTC, down from $90,000 at the start of the year following the difficulty drop. But this average conceals enormous dispersion:
| Miner Tier | Electricity Cost | Hardware Efficiency | Est. Cost per BTC | |-----------|-----------------|--------------------|--------------------| | Tier 1 (Top decile) | $0.03–0.04/kWh | <15 J/TH (S21+) | $34,000–$43,000 | | Tier 2 (Median) | $0.05–0.06/kWh | 15–20 J/TH | $55,000–$70,000 | | Tier 3 (Marginal) | $0.07–0.08/kWh | 20–25 J/TH | $75,000–$90,000 |
With Bitcoin trading in the $75,000–$100,000 range for much of early 2026, Tier 3 miners are operating at or below breakeven. In 2026, mining a single Bitcoin requires approximately 854,400 kWh of electricity globally — a figure that has risen roughly 30% since the April 2024 halving due to difficulty increases.
The breakeven hashcost at a power rate of $0.05/kWh is approximately $38.40 per PH/s per day. At the current hashprice of $23.9, even mid-efficiency miners are underwater on a pure electricity basis before accounting for depreciation, labor, or debt service. Mining ROI timelines have stretched past 1,000 days for new hardware deployments, compared to roughly 300–400 days in the pre-halving era.
Winter Storm Fern, which struck the U.S. in late January 2026, exposed a critical vulnerability: the extreme geographic concentration of Bitcoin hashrate in the ERCOT (Texas) grid region. During peak disruption, ERCOT North Hub electricity prices spiked to approximately $1,200 per MWh — roughly 20x the normal rate — making continued mining economically irrational.
Miners responded rationally. Large-scale operators including Marathon (MARA), Riot Platforms, and CleanSpark curtailed operations proactively, both to preserve margins and to participate in demand-response programs that pay miners to shed load during grid emergencies. But the macro impact was severe: 455 EH/s went temporarily offline across the network, representing roughly 40% of the global hashrate at the time.
The episode revived a debate about whether Bitcoin's proof-of-work security model can tolerate such concentration risk. If 40% of global hashrate can be disrupted by a single regional weather event, the network's censorship resistance is weaker than the zetahash headline suggests. Notably, mining operations in the UAE, Kazakhstan, and Paraguay experienced no disruption and absorbed a temporarily larger share of block production during the outage window.
The most consequential development in Bitcoin mining in 2026 is not happening on the Bitcoin network at all. It is happening inside the data centers that miners already own.
A wave of public mining companies have announced or executed pivots toward AI inference and high-performance computing, leveraging their core competitive advantage: access to large-scale, low-cost power infrastructure. The numbers are staggering:
By Q4 2025, AI and HPC revenue already accounted for 10–20% of total revenue for diversified miners. Industry projections suggest mining revenue could fall from approximately 85% of total revenue in early 2025 to less than 20% by late 2026 for companies that have secured AI contracts.
The economics are clear. The cost to build AI-ready facilities has escalated to approximately $8–11 million per MW, driven by liquid cooling requirements and transformer shortages. But miners who already own permitted, grid-connected sites with 100+ MW of capacity hold assets that hyperscalers desperately need — and are willing to pay a premium for.
This is a one-way door. Once a facility is retrofitted from ASIC racks to GPU clusters with liquid cooling, the economics of switching back to Bitcoin mining become prohibitive. The industry is undergoing an irreversible structural transformation.
As margins compress, consolidation is accelerating. In February 2026, American Bitcoin Corp. — a venture launched by Hut 8 in partnership with Eric Trump — announced a definitive merger agreement with Gryphon Digital Mining (GRYP) in a stock-for-stock transaction. The combined entity will operate under the American Bitcoin brand and trade on Nasdaq as "ABTC," with the stated ambition of becoming the world's largest pure-play Bitcoin miner while building a strategic Bitcoin reserve.
Galaxy Digital's research division has flagged a growing M&A pipeline, predicting that mergers, acquisitions, and outright failures will consolidate hashrate among a smaller group of hybrid operators with both mining and AI/HPC revenue streams.
The year-to-date stock performance of the top three public miners tells the story of divergence:
| Miner | YTD Performance (Feb 2026) | Strategy | |-------|---------------------------|----------| | CleanSpark (CLSK) | +53% | Efficient pure-play; low-cost operations | | Riot Platforms (RIOT) | +48% | Hybrid; Texas demand-response revenue | | Marathon (MARA) | -28% | Largest hashrate; high BTC treasury exposure |
Marathon's underperformance — despite holding a top-10 public Bitcoin treasury — illustrates the market's current pricing logic: operational efficiency and revenue diversification matter more than raw hashrate or BTC holdings.
The emerging survival criteria for Bitcoin miners in 2026 are clear and unforgiving:
Miners who check all five boxes — Hut 8, Core Scientific, TeraWulf, IREN — are being re-rated by the market as energy infrastructure companies that happen to mine Bitcoin. Those who check one or two are acquisition targets. Those who check none are on borrowed time.
The Bitcoin mining industry's zetahash milestone was supposed to be a triumph. Instead, it marks the beginning of an identity crisis. The network has never been more computationally powerful — and the business of securing it has never been less profitable on a per-unit basis.
The companies that will define the next chapter of this industry are not the ones deploying the most hashrate. They are the ones that recognized earliest that their real asset was never the ASIC — it was the megawatt. The $7 billion AI deals, the AWS leases, the Anthropic partnerships: these are not sidelines. They are the new main event.
For Bitcoin maximalists, this raises an uncomfortable question. If the economic incentive for miners is increasingly to not mine Bitcoin, what does that mean for the long-term security budget of the network? The halving cycle was designed to transition security funding from block subsidies to transaction fees. But with fee revenue still a fraction of block rewards, and miners finding far more lucrative uses for their power infrastructure, the gap between Bitcoin's security needs and its economic incentives is widening.
The miners who survive 2026 will be those who treated Bitcoin mining not as an end in itself, but as an on-ramp to becoming the energy infrastructure backbone of the AI era. The rest will be acquired, converted, or switched off.