Mining farm
A Bitcoin mining farm is a facility that operates many ASIC miners with shared electrical, cooling, networking and monitoring infrastructure. Farms range from small rooms to industrial sites with loads measured in tens or hundreds of megawatts. They perform the same proof-of-work function as an individual miner; scale changes the engineering and economics, not Bitcoin's consensus rules.
The term is also used for facilities mining other proof-of-work cryptocurrencies. Hardware, algorithms and operating requirements differ, so a farm built for Bitcoin's SHA-256 ASICs cannot be assumed to mine another network efficiently.
Development and ownership models
Early Bitcoin participants could mine on personal computers. GPUs, FPGAs and then purpose-built SHA-256 ASICs increased both performance and power density. As ASIC mining became specialized, operators concentrated machines in locations where they could obtain adequate electrical service, cooling, network access and maintenance. Containerized modules made it possible to deploy a repeatable unit near an energy source, although a container does not remove the need for substations, permits or trained staff.
A farm's operator, hardware owner and hashrate customer can be different parties. Common arrangements include owner-operated facilities, colocation or hosting for customers' machines, joint ventures and publicly listed mining companies. A mining pool coordinates jobs and payouts but does not necessarily own the facilities whose hashrate connects to it. Cloud mining may represent rented hashrate or only a contractual financial claim and should not be treated as proof that the customer owns identifiable equipment.
Equipment and operation
Each ASIC repeatedly hashes candidate block headers. A farm supplies work through a pool connection or its own mining stack, collects shares and monitors accepted work, hardware errors and downtime. A local Bitcoin node may validate transactions and blocks or provide templates, while fleet software controls device settings and alerts.
Major facility systems include:
- high-voltage utility service, transformers, switchgear and power-distribution units;
- racks or containers holding ASIC miners and network equipment;
- air, evaporative or liquid/immersion cooling;
- fire detection, physical security and spare-parts inventory;
- monitoring for power, temperature, humidity, hashrate and failed fans or boards; and
- network links to pools, nodes and management systems.
Nearly all electricity entering the miners becomes heat. Air-cooled farms move large volumes of filtered outside air and must manage dust, noise and recirculation. Evaporative systems can reduce temperature but consume water and depend on local humidity. Immersion systems place equipment in dielectric fluid and move heat through exchangers; they can support high power density but add fluid, pump and maintenance requirements. Some sites reuse heat for buildings, greenhouses or industrial processes, but useful recovery depends on temperature, distance and a matching local demand.
Availability matters as much as nameplate capacity. Power interruptions, pool latency, failed boards, thermal throttling and slow repairs reduce effective hashrate. Operators therefore track uptime, joules per terahash, accepted shares and energy use rather than relying only on the manufacturer's advertised rate.
Hashrate, difficulty and revenue
A farm's hashrate is the number of SHA-256 attempts it performs per second. More hashrate increases its expected share of blocks or pool payouts, but does not solve a fixed queue of progressively harder puzzles. Bitcoin adjusts mining difficulty every 2,016 blocks so that blocks continue to arrive at approximately the protocol's target rate despite changes in total network hashrate.
Revenue comes from the block subsidy and transaction fees allocated under the farm's solo or pool arrangement. A pool can reduce variance by paying for shares before one participant finds a full block, in exchange for fees and counterparty or protocol risk. Costs include electricity, cooling, staffing, rent, finance, pool fees, repairs, insurance, taxes and hardware depreciation. Profitability changes with difficulty, fees, the subsidy, uptime and bitcoin's exchange rate.
Hardware efficiency is usually expressed in joules per terahash (J/TH); a lower figure means less electrical energy per unit of hashing. The Cambridge Centre for Alternative Finance estimated the network's average efficiency at 28.2 J/TH in June 2024, a 24% improvement from the previous year.[1] A more efficient machine may still be uneconomic if acquisition, hosting or financing costs are too high. Likewise, a dated facility hashrate or share of global mining quickly becomes obsolete and should not be treated as a permanent ranking.
Electricity use and emissions
Bitcoin mining is electricity-intensive, but global totals and energy sources are modelled estimates rather than a complete census. The 2025 Cambridge report surveyed 49 firms representing nearly 48% of the network's implied hashrate. It estimated annual use at 138 TWh, about 0.54% of global electricity consumption, and modelled emissions of 39.8 million tonnes of carbon-dioxide equivalent. It reported a 52.4% sustainable energy share, comprising 42.6% renewable and 9.8% nuclear energy, for the study period. Electricity accounted for more than 80% of respondents' cash operating expenditure.[1]
Those figures should be read with the report's sampling and modelling limits. Its respondents were concentrated in North America, so their locations and electricity contracts do not describe every miner. Renewable, carbon-free and sustainable are also not interchangeable terms.
Local impact depends on the marginal electricity source, grid congestion, time of consumption, noise, water use and air pollution from generation. A study of 34 large United States facilities estimated that 85% of the additional electricity associated with them during its August 2022–July 2023 study period came from fossil-fuel generation, illustrating why annual contractual energy shares alone do not settle emissions or local air-pollution questions.[2]
Flexible load and grid interaction
ASICs can stop and restart more quickly than many industrial processes. Some farms therefore participate in utility demand-response programmes or curtail when electricity prices rise. This can give a grid operator controllable load and can improve the economics of generation that would otherwise be curtailed. It can also create a large new baseline demand, so the result depends on market rules, location and what generators respond.
Economic modelling of the Texas electricity market found that mining could support additional renewable capacity while also increasing emissions; demand response reduced but did not automatically eliminate that trade-off.[3] Payments for curtailment should be described separately from block rewards and ordinary energy purchases.
Hardware life cycle and electronic waste
ASICs become uncompetitive when more efficient models or higher difficulty make their operating cost exceed expected revenue. They may be resold to locations with cheaper electricity, used for spare parts or recycled rather than discarded immediately. The Cambridge report estimated 2.3 kilotonnes of electronic waste from Bitcoin mining for its study period and reported that 86.9% of decommissioned hardware in its respondent sample was recycled or repurposed.[1] These results are estimates and survey responses, not a registry of every machine.
Useful-life estimates must distinguish technical failure from economic retirement. A working older ASIC can return to service if electricity prices or market conditions change, but storage, transport and declining efficiency still have costs.
Geography and regulation
Mining locations change as electricity prices, regulation, taxes, capital access and hardware logistics change. China was widely reported as the dominant location before restrictions announced in 2021 shifted substantial hashrate to other countries. Older claims that a fixed percentage of all mining permanently occurs in China, Iceland, Georgia or another jurisdiction are historical snapshots.
Industrial projects may require land-use, electrical-interconnection, noise, environmental and construction approvals. Rules can apply to energy consumption, tax incentives, financial reporting or participation in electricity markets. Pool traffic and internet addresses provide incomplete geographic evidence because machines can use proxies and pool operators do not necessarily disclose their customers.
Home and hosted mining
A home installation uses the same specialized ASIC principle at smaller scale. Practical limits include circuit capacity, noise, exhaust heat, ventilation and residential electricity prices. Heat can be useful in cold weather, but it does not make the electricity free and can be a liability in summer.
Hosting transfers some facility work to an operator but introduces counterparty and contract risk. Prospective customers need to distinguish ownership of a specific machine from a promise of hashrate, identify electricity and repair charges, and understand termination or relocation rights. A dashboard is not independent evidence that named equipment exists or remains profitable.
See also
References
- ↑ 1.0 1.1 1.2 Cambridge Centre for Alternative Finance, Cambridge Digital Mining Industry Report: Global Operations, Sentiment, and Energy Use, April 2025.
- ↑ Gianluca Guidi et al., “The environmental burden of the United States’ bitcoin mining boom”, Nature Communications, 26 March 2025.
- ↑ Alex Bruno, Philip Weber and Andrew J. Yates, “Can Bitcoin mining increase renewable electricity capacity?”, Resource and Energy Economics, volume 74, 2023.