Bitcoin's network hashrate sits at 923 EH/s as of October 8, 2026 — 17% below the all-time high of 1,110 EH/s recorded in November 2025. Network difficulty has fallen 19.3% from its peak of 155.97 trillion to 132.76 trillion. The hashprice, the single metric that captures miner revenue per unit o...
"Mining revenue is projected to plummet from around 85% of total revenue in early 2025 to less than 20% by the end of 2026 for companies that have secured AI contracts." — CoinShares, Q2 2026 Bitcoin Mining Report
Bitcoin's network hashrate sits at 923 EH/s as of October 8, 2026 — 17% below the all-time high of 1,110 EH/s recorded in November 2025. Network difficulty has fallen 19.3% from its peak of 155.97 trillion to 132.76 trillion. The hashprice, the single metric that captures miner revenue per unit of computation, bottomed at $29 per PH/s per day in Q1 2026, the lowest level since the April 2024 halving cut the block subsidy from 6.25 BTC to 3.125 BTC.
Simultaneously, publicly listed mining companies have signed over $32 billion in aggregate AI and high-performance computing (HPC) contracts, redirecting power capacity away from Bitcoin hashing and toward GPU-based workloads. TeraWulf alone has locked $12.8 billion in HPC lease agreements. The result is a structural divergence: the companies that once secured Bitcoin's network are becoming data-center landlords, and the network's security budget is contracting in dollar terms even as Bitcoin trades near $83,000.
Mining stocks have underperformed Bitcoin itself. Over the past twelve months, CleanSpark (CLSK) is down 40%, MARA Holdings (MARA) is down 51%, while the iShares Bitcoin Trust ETF declined 33% over the same period. The market is repricing these firms not as Bitcoin proxies but as infrastructure operators — and discounting the mining side of the business.
Bitcoin crossed the 1 ZH/s (1,000 EH/s) threshold for the first time in late 2025. The network peaked at approximately 1,110 EH/s in November 2025, according to Hashrate Index data. That number did not hold.
A severe U.S. winter storm in January 2026 forced Texas-based miners to shut down rigs to protect the power grid, triggering a 12% drawdown — the largest since China's 2021 mining ban. But while weather was the proximate cause, the structural cause was economic: the halving had cut the block subsidy in half, and Bitcoin's price had begun a protracted decline from its all-time high above $126,000 in October 2025.
By mid-2026, the average network hashrate had fallen to a range of 740–886 EH/s, according to CoinShares' Q2 2026 mining report. The network recorded ten downward difficulty adjustments against just seven increases through mid-year. In June, difficulty dropped 10.09% to 124.93 trillion — the 11th-largest downward adjustment in the network's history and the second-largest of 2026.
As of October 8, 2026, the seven-day average hashrate stands at 923 EH/s (CoinWarz, block 970,520), and difficulty sits at 132.76 trillion after a modest +0.38% retarget on October 3. The network has stabilized but has not returned to its zettahash-era peak.
Hashprice — the dollar revenue a miner earns per petahash per second per day — is the sector's single most important operating metric. It fell to $29/PH/s/day in Q1 2026, a five-year low, according to CoinShares. By August 2, hashprice had only partially recovered to $31.80/PH/s/day.
As of early October 2026, hashprice has rebounded to approximately $40/PH/s/day, carried by Bitcoin's partial price recovery to the $83,000–$85,000 range and the 19% reduction in difficulty that lowered the competition per hash. This represents a roughly 38% recovery from the trough, but remains well below the $55–$65/PH/s/day levels that prevailed through most of 2025.
Production costs tell the rest of the story. CoinShares' Q2 2026 report pegged the weighted average cash cost to produce one bitcoin among listed miners at approximately $75,500. Bitcoin closed Q2 at $58,400, meaning the listed mining sector operated below cash breakeven in aggregate for the quarter.
By comparison, CoinShares estimated Q4 2025 weighted average cash costs at approximately $79,995/BTC, when Bitcoin was still trading above $100,000. The numbers reveal a sector where costs have been slow to decline even as revenue has fallen sharply.
The variation among individual operators is extreme. According to CoinShares data, production costs ranged from $871/BTC at ABTC (which benefits from unique power arrangements) to $134,446/BTC at TeraWulf — a reflection of TeraWulf's shift in focus toward AI infrastructure rather than optimizing for Bitcoin production. CleanSpark reported cash costs of $71,995/BTC; Riot Platforms came in at $74,955/BTC.
According to CoinShares, roughly 15%–20% of the global Bitcoin mining fleet was unprofitable at Q1-Q2 2026 hashprice levels, particularly operators running older hardware or facing electricity costs above $0.08/kWh.
The economics driving the AI pivot are straightforward: a megawatt of power allocated to GPU-based AI training or inference can generate several times more revenue than a megawatt allocated to Bitcoin hashing. Mining firms already own the three assets that AI companies need — power interconnections, cooling infrastructure, and physical space — and they own them in locations where new grid capacity is increasingly scarce.
The aggregate value of disclosed AI/HPC contracts across publicly listed Bitcoin miners now exceeds $32 billion:
The industry has adopted what some operators call the "Mullet Data Center Strategy": AI workloads in the front (stable, contracted, high-margin), Bitcoin mining in the back (flexible, interruptible, used to balance grid demand and absorb excess power).
CoinShares projects that for companies with AI contracts, mining revenue will fall from approximately 85% of total revenue in early 2025 to less than 20% by end-2026. These are not mining companies that also do AI. They are becoming AI infrastructure companies that also mine Bitcoin.
The most recent quarterly filings illustrate the bifurcation:
Marathon Digital (MARA): Q3 revenue of $252 million and $123 million in net income, mining 2,144 BTC. Marathon reported a 64% year-over-year revenue increase, driven primarily by its expanded hashrate (above 36 EH/s) and Bitcoin's partial price recovery. Marathon has been slower to pivot toward AI relative to peers.
Riot Platforms (RIOT): Q3 revenue of $180 million and $104 million in net income. Riot operates above 36 EH/s of deployed hashrate. The company has announced AI/HPC strategy plans but has not yet disclosed contract values comparable to TeraWulf or Core Scientific.
CleanSpark (CLSK): Revenue fell 30.5% year-over-year to $138 million in fiscal Q3 2026. The company swung to a net loss of $239.8 million from net income of $257.4 million in the same quarter of the prior year. CleanSpark has focused on mining efficiency rather than a large-scale AI pivot.
TeraWulf (WULF): Q2 revenue of $44.8 million, with HPC lease revenue representing 71% of the total. Q1 revenue was $34.0 million, including $21.0 million from HPC. TeraWulf's per-bitcoin mining cost of $134,446 reflects its strategic de-emphasis of mining operations.
The data reveals two distinct strategies: Marathon and Riot are doubling down on hashrate and mining scale; TeraWulf, Core Scientific, and Hut 8 are converting to AI infrastructure with mining as a secondary, flexible workload. The market has so far rewarded neither approach — all mining stocks are down substantially year-over-year.
As of October 8, 2026:
| Ticker | Price | 1-Day Change | 1-Year Change | |--------|-------|-------------|---------------| | CLSK | $10.72 | -7.0% | -40% | | MARA | $9.89 | -5.0% | -51% | | IBIT (BTC ETF) | — | -2.7% | -33% |
Bitcoin ETFs recorded $484.9 million in net outflows on October 7; Ethereum ETFs saw $160.9 million in outflows. The macro environment — rising oil prices from Middle East tensions, persistent inflation — is pressuring risk assets broadly, but mining equities are absorbing disproportionate losses.
The underperformance of mining stocks relative to Bitcoin itself is a structural phenomenon, not a one-day event. Over the past year, every major listed miner has trailed the underlying commodity. The market is pricing in dilution from share issuance (public miners sold 32,000 BTC and issued equity in Q1 2026 alone, according to industry data), rising power costs, and uncertainty about whether AI revenue will materialize at scale.
The 17% decline in hashrate from peak raises questions about Bitcoin's security budget. The network's annual security spend — block subsidies plus transaction fees, denominated in dollars — has contracted as both Bitcoin's price and the number of active miners have declined.
At 923 EH/s and a block subsidy of 3.125 BTC (approximately $260,000 per block at current prices), the network pays roughly $37.4 million per day to miners in subsidy alone. Transaction fees add a variable component, but fee revenue has been modest relative to the subsidy since the Ordinals-driven fee spike faded.
The deeper question is what happens as more mining capacity is redirected to AI workloads. If the AI pivot continues at its current pace, the Bitcoin network may need to rely on a smaller cohort of dedicated mining firms and solo operators for its security. The 19% difficulty decline has made mining more accessible for smaller operators with low power costs — effectively redistributing hashrate from publicly listed firms to private miners in low-cost jurisdictions.
Cambridge's April 2025 industry report estimated Bitcoin's annual electricity consumption at approximately 138 TWh, or 0.5% of global electricity use. Of that, 52.4% came from sustainable sources: 42.6% renewables and 9.8% nuclear. As large miners redirect power to AI, the network's absolute energy consumption is likely declining, though no updated Cambridge estimate is available for Q4 2026.
The Bitcoin mining sector in October 2026 is undergoing a structural transformation that has no precedent in the network's seventeen-year history. The April 2024 halving compressed margins. The subsequent price decline from $126,000 to the low-$80,000 range compressed them further. And the AI infrastructure boom offered an exit ramp that a critical mass of publicly listed miners have taken.
The result is a network that is less centralized among public companies, potentially more reliant on private miners in low-cost jurisdictions, and generating less total security spend in dollar terms. For investors, the mining sector is no longer a leveraged Bitcoin bet — it is an infrastructure allocation decision where the marginal dollar of capital is flowing toward GPU racks, not ASIC rigs.
The firms that remain committed to mining — Marathon, Riot, CleanSpark — are betting that Bitcoin's price will recover sufficiently to restore margins. The firms that have pivoted — TeraWulf, Core Scientific, Hut 8 — are betting that contracted AI revenue is worth more than optionality on Bitcoin's price. The market, as of this writing, is skeptical of both strategies.