Bitcoin's network hashrate crossed 1,000 exahashes per second on October 13, 2026, reclaiming the one-zettahash threshold eight months after a 27% decline that bottomed near 850 EH/s in February. Three days later, on October 16, mining difficulty adjusted upward by 5.92% to 140.58 trillion at blo...
"Q2 2026 pushed the listed mining sector below cash breakeven in aggregate." — CoinShares, Bitcoin Mining Report Q2 2026
Bitcoin's network hashrate crossed 1,000 exahashes per second on October 13, 2026, reclaiming the one-zettahash threshold eight months after a 27% decline that bottomed near 850 EH/s in February. Three days later, on October 16, mining difficulty adjusted upward by 5.92% to 140.58 trillion at block 971,712 — the third-largest positive adjustment of the year and the highest reading since March 5.
The recovery masks a structurally different mining economy. CoinShares' Q2 2026 report pegged the weighted average cash cost of producing one bitcoin at $75,500, well above the quarter-end spot price of roughly $58,400. Transaction fees contributed less than 1% of block revenue, down from 5–7% before the April 2024 halving. Miners that survived did so by upgrading to sub-15 J/TH hardware, locking in sub-$0.08/kWh power, or pivoting capacity toward AI and high-performance computing contracts that now account for over $70 billion in cumulative deal value across the publicly listed sector.
The hashrate milestone is a measure of raw computational commitment, not profitability. What the data shows is a network secured by an unprecedented volume of silicon, financed increasingly by revenue streams that have nothing to do with Bitcoin block rewards.
The Bitcoin network's seven-day average hashrate touched 1,024.96 EH/s on October 16, 2026, according to Hashrate Index data. The weekly average subsequently settled at 1,031.61 EH/s. One zettahash equals one sextillion (10²¹) hash computations per second.
The milestone does not alter Bitcoin's consensus rules, block timing, or security model in any mechanical sense — difficulty adjustments ensure roughly 10-minute block intervals regardless of hashrate level. It is, however, a proxy for the aggregate capital deployed into mining infrastructure globally. Publicly traded miners alone operate approximately 413.8 EH/s, representing about 42% of total network compute, led by Marathon Digital Holdings at 70.3 EH/s, Riot Platforms at 37.2 EH/s, and CleanSpark at 36.0 EH/s.
The path to this point was not linear. Bitcoin's hashrate peaked at approximately 1,160 EH/s in October 2025 before declining to 850 EH/s by early February 2026 — a 27% peak-to-trough drop, the steepest sustained decline since China's mining ban in mid-2021.
Three factors converged to cause the collapse:
The recovery began in Q2 2026 as Bitcoin's price stabilized and then climbed above $80,000. By October 13, the network had regained the 1 ZH/s level. The hashrate rebounded to 1,075 EH/s that day before the difficulty adjustment on October 16 repriced the competitive landscape.
CoinShares' Q2 2026 Bitcoin Mining Report, covering publicly listed miners, found that the weighted average ex-tax cash cost to produce one bitcoin reached approximately $75,500. Bitcoin's price at the end of Q2 was roughly $58,400, meaning the aggregate listed mining sector operated below cash breakeven.
Cost variation across operators was wide:
By October, conditions had improved. Hashprice — the dollar revenue per petahash per day — climbed to $39.33 in September, the strongest monthly average since January 2026, briefly topping $40. With BTC trading near $82,000 on October 18, the economics have shifted materially in miners' favor. Network-wide monthly revenue reached approximately $1.12 billion in the most recent reporting period, up 11% from August.
The October 16 difficulty adjustment to 140.58 trillion reflected faster-than-target block production as hashrate surged. At 5.92%, it was the third-largest upward adjustment of 2026, activated at block 971,712 at 12:37:31 UTC. The reading is the highest since March 5 but sits approximately 15% below the all-time record of 155.97 trillion set on October 29, 2025.
The current block subsidy is 3.125 BTC per block. Total block reward fluctuates between 3.1 and 3.4 BTC depending on mempool congestion, with transaction fees contributing a persistently small share. The next halving, which will reduce the subsidy to 1.5625 BTC, is expected in April 2028.
Hardware efficiency is the primary lever miners can pull to survive in a post-halving, high-difficulty environment. The current generation of application-specific integrated circuits (ASICs) represents a step-change from prior models:
| Model | Hashrate | Efficiency | Cooling | |---|---|---|---| | Bitmain Antminer S23 Hyd 3U | — | 9.5 J/TH | Hydro | | Bitmain Antminer S21 XP Hyd | — | 12.0 J/TH | Hydro | | Bitmain Antminer S21 XP | 270 TH/s | 13.5 J/TH | Air | | Bitmain Antminer S21 Immersion | 215 TH/s | 14.0 J/TH | Immersion | | Bitmain Antminer S21 Pro | 234 TH/s | 15.0 J/TH | Air |
For context, the Antminer S19 XP, which was a flagship model until recently, rated at 21.5 J/TH. The S21 XP at 13.5 J/TH produces equivalent hashrate at roughly 37% less electricity. Industrial miners running sub-15 J/TH hardware at $0.07/kWh hosting report 42–60% operating margins at current BTC prices, according to industry estimates.
The efficiency gap between frontier hardware and older-generation machines is widening. An S21 XP earns the same bitcoin as two Antminer S9s at approximately one-tenth the electricity cost, effectively rendering pre-2023 hardware nonviable at any electricity rate above $0.04/kWh.
Hashrate Index's Q3 2026 data shows the following estimated country-level distribution:
| Country | Share | Est. Hashrate | |---|---|---| | United States | 35.6% | ~335 EH/s | | Russia | 18.1% | ~170 EH/s | | China | 11.7% | ~110 EH/s | | Paraguay | 4.7% | ~44 EH/s | | UAE | 3.0% | ~28 EH/s | | Oman | 2.8% | ~26 EH/s | | Canada | 2.6% | ~24 EH/s |
Three countries — the U.S., Russia, and China — control an estimated 65.4% of global hashrate. The U.S. share has declined from approximately 41% a year ago to 35.6%, shedding roughly 54 EH/s, attributed partly to Texas curtailments and miners reallocating capacity to AI workloads. Russia's share, at 18.1%, is the highest level recorded in Hashrate Index's dataset. China retains approximately 12% through smaller and unofficial operations despite the 2021 ban. Kazakhstan's share has decreased due to regulatory tightening.
The Gulf region — UAE and Oman combined at 5.8% — represents a growing hub driven by state-backed energy subsidies and sovereign fund investments in mining infrastructure.
The most consequential structural shift in the mining sector is the reallocation of power capacity toward AI and high-performance computing. Over $70 billion in cumulative AI/HPC contracts have been announced across the public mining sector.
The economics explain the move. HIVE Digital Technologies has estimated that 10 MW of Nvidia H100 GPU infrastructure can produce revenue comparable to 100 MW of Bitcoin mining capacity — a 10:1 power-to-revenue advantage. CoinShares projects that listed miners with AI contracts could derive up to 70% of their revenue from AI by year-end 2026, up from roughly 30% at the start of the year.
The valuation gap is stark. Miners with secured HPC contracts trade at 12.3x expected next-twelve-month revenues. Pure-play Bitcoin miners trade at 5.9x. At least eight publicly listed mining firms — including IREN, Core Scientific, Riot Platforms, TeraWulf, CleanSpark, Marathon Digital, and Bitfarms — have announced partial or full transitions into AI/HPC hosting.
This creates a paradox for Bitcoin's security model. The network's hashrate is being subsidized, in part, by AI revenue. Miners keep their Bitcoin operations running at marginal or below-marginal returns because the incremental electricity cost is low once data center infrastructure is built for AI workloads. If AI demand contracts or power costs shift, the implicit subsidy to Bitcoin security could unwind.
The Cambridge Bitcoin Electricity Consumption Index estimates the network's annualized draw at 170–180 TWh. Digiconomist's estimate runs higher at 204 TWh. The discrepancy reflects methodological differences in estimating hardware fleet composition and operational uptime.
Cambridge's most recent industry survey indicates 52.4% of mining electricity comes from sustainable sources: 42.6% renewables and 9.8% nuclear. These figures are self-reported by miners and have been contested by environmental researchers who note the reliance on accounting technicalities such as renewable energy certificates rather than direct consumption of clean power.
At 1 ZH/s, the network consumes roughly 170–200 TWh to process approximately 7 transactions per second on the base layer — an efficiency metric that proponents argue is irrelevant (Bitcoin's value proposition is settlement finality, not throughput) and critics cite as evidence of structural waste.
The one-zettahash milestone is a data point, not an inflection point. Bitcoin's mining network has recovered from its worst downturn since the China ban, but the structure of that recovery reveals a sector that is financing security through an increasingly indirect economic model. AI revenue subsidizes hashrate. Hardware efficiency gains compress the field to a handful of well-capitalized operators. Geographic concentration persists, with three countries controlling two-thirds of compute.
The April 2028 halving will cut the block subsidy to 1.5625 BTC — roughly $128,000 in annual revenue per block at current prices. Transaction fees show no trend toward filling the gap. The question that the mining industry has deferred, not answered, is whether Bitcoin's security budget can sustain itself from transaction demand alone once subsidies become negligible. At 1 ZH/s, the network has never been more computationally secure. The economics underwriting that security have never been more dependent on non-Bitcoin revenue.