Bitcoin Mining's Energy Mix in 2026: Renewables, Stranded Gas, and Grid Balancing

Bitcoin mining's energy mix has shifted dramatically since 2021. The industry that once drew over a third of its electricity from coal now sources more than half from sustainable generation. At the same time, miners have become significant participants in grid balancing programs, offering gigawatts of flexible load that can curtail in seconds. This article examines the data behind these claims, the strategies miners use to source power, and the criticisms that remain valid in 2026.

Where Bitcoin Mining Gets Its Power

The most rigorous data on mining's energy mix comes from the Cambridge Centre for Alternative Finance (CCAF), which published its Digital Mining Industry Report in April 2025. The study surveyed 49 mining companies across 16 jurisdictions, representing 48% of global hashrate.

The headline finding: 52.4% of Bitcoin mining electricity now comes from zero-emission sources, up from 37.6% in 2022. That breaks down into 42.6% from renewables (hydropower 23.4%, wind 15.4%, solar 3.2%) and 9.8% from nuclear. Natural gas has become the single largest individual source at 38.2%, while coal has dropped to 8.9%, down from 36.6% in 2022.

The Bitcoin Mining Council's Q4 2025 survey, which covers roughly half of global hashrate through voluntary self-reporting, broadly corroborates these numbers with a reported 50-60% sustainable power mix. However, BMC data is self-reported by member companies and not independently audited, so the Cambridge figures are generally treated as more authoritative.

Methodological note: Neither CCAF nor BMC data covers the full network. The CCAF sample represents 48% of global hashrate, and the unsampled portion (concentrated in Russia, China, and Central Asia) likely uses more fossil fuels. Treat the 52.4% sustainable figure as a floor estimate for the sampled population, not a definitive global number.

How Much Energy Does Bitcoin Mining Use?

The Cambridge Bitcoin Electricity Consumption Index (CBECI) estimates Bitcoin's annual electricity consumption at approximately 138 TWh as of early 2025. The network's estimated emissions stand at 39.8 MtCO2e. For context, 138 TWh is roughly comparable to the annual electricity consumption of Poland or Argentina, and represents approximately 0.5% of global electricity production.

Total consumption has risen as the network hashrate crossed the 1 zettahash per second (ZH/s) milestone in September 2025, a roughly 25% increase from the start of that year. But total consumption tells only part of the story. What matters increasingly is where that energy comes from, whether it would otherwise be wasted, and how flexibly miners consume it.

Geographic Distribution After the China Ban

China's 2021 mining ban reshuffled the global hashrate map. The United States absorbed the largest share of displaced miners and now accounts for roughly 37.5% of global hashrate. Russia holds approximately 16.4%, while China itself retains an estimated 11.7% through semi-tolerated operations, many leveraging seasonal Sichuan hydropower.

Several trends stand out. Kazakhstan, which briefly surged to roughly 18% of hashrate after the China ban, has declined to just 2.1% due to energy caps and regulatory pressure. Meanwhile, new hydro-powered hubs have emerged: Paraguay grew 54% year-over-year with electricity costs as low as $0.033/kWh from Itaipu Dam surplus. Ethiopia, powered by the Grand Ethiopian Renaissance Dam (inaugurated September 2025), now hosts roughly 23 mining operations consuming approximately 600 MW. Both countries face regulatory uncertainty, with proposed tariff increases that could reshape their mining industries by 2027.

Stranded Gas: Turning Waste Into Hashrate

One of Bitcoin mining's most compelling sustainability arguments involves stranded and flare gas capture. When oil wells produce natural gas as a byproduct in locations far from pipelines, operators typically flare (burn) or vent it. The World Bank's 2025 Global Gas Flaring Tracker reported that global flaring rose to 151 billion cubic meters in 2024, the highest level since 2007, releasing 389 million tonnes of CO2 equivalent.

Bitcoin miners have found an economic use for this otherwise wasted energy. Crusoe Energy became the most prominent example, deploying over 425 modular data centers across seven U.S. states and Argentina. The company captured nearly 22 billion cubic feet of natural gas that would have been flared, mitigating 2.7 million metric tons of greenhouse gas emissions. In 2024 alone, Crusoe converted 10 billion cubic feet of captured gas into 1.3 TWh of power for mining operations.

Notably, Crusoe sold its Bitcoin mining business to NYDIG in early 2025 to focus on AI infrastructure, forecasting $998 million in 2025 revenue. The pivot illustrates a broader trend where mining infrastructure becomes a gateway to higher-margin compute workloads. But the environmental model remains: gas-to-compute operations reduce CO2-equivalent emissions by up to 63% compared to traditional flaring, because enclosed gas engines combust methane more completely than open-air flares.

Other companies continue this approach. Giga Energy focuses on flare gas capture in Texas, while EZ Blockchain deploys modular “SmartGrid” systems directly on well pads, in some cases reducing flaring from 240,000 standard cubic feet per day to zero.

Hydroelectric Surplus Mining

Hydroelectric power represents 23.4% of Bitcoin mining's energy mix, making it the single largest renewable source. This concentration reflects a deliberate strategy: miners seek out regions with hydroelectric surplus where generation exceeds local demand and transmission capacity.

Paraguay and the Itaipu surplus

Paraguay generates far more hydroelectricity than it consumes domestically, primarily from the Itaipu Dam (shared with Brazil) and the Yacyretá Dam. Mining operations have grown 54% year-over-year, with HIVE Digital Technologies acquiring Bitfarms' Yguazu facility (200 MW) for $56 million in January 2026. However, ANDE, the national utility, has warned that uncontrolled mining expansion could jeopardize national power stability, and proposed tariff increases may make roughly half of current operations unprofitable by mid-2026.

Quebec and Nordic hydro

Quebec's grid is almost entirely carbon-free, offering some of the lowest electricity rates globally. Canada holds approximately 2.6% of global hashrate. In the Nordic countries, Iceland leverages geothermal energy for stable baseload mining power, while Norway, despite its vast hydroelectric resources, shifted policy in 2025 against new mining facilities, signaling that even green energy faces allocation conflicts when competing with other industrial needs.

Ethiopia and the GERD

Ethiopia has emerged as an unexpected mining hub, with roughly 2.6% of global hashrate powered by the Grand Ethiopian Renaissance Dam. The country generated approximately $220 million from mining concessions in 2024. However, a freeze on new crypto mining permits and planned tariff increases of 30% or more, starting in late 2025, threaten this trajectory.

Grid Balancing: Miners as Flexible Load

Perhaps the strongest structural argument for Bitcoin mining's energy role is its function as controllable load. Unlike most industrial consumers, mining operations can curtail to near-zero power draw within seconds, without equipment damage or lost product. This makes miners uniquely suited for demand response programs where grid operators pay large consumers to reduce load during peak demand periods.

The Texas model

Texas's ERCOT grid has become the primary testbed for miner-as-flexible-load. As of November 2025, crypto mining electric demand in Texas reached 4,288 MW, with projections to surpass 5,300 MW by 2027. ERCOT approved approximately 9,500 MW of large flexible load capacity by the end of 2025, with this category (which includes AI data centers) expected to consume 54 billion kWh in 2025, roughly 10% of total ERCOT forecast consumption.

The economics are concrete. Riot Platforms, operating North America's largest Bitcoin mining facility at 700 MW in Rockdale, Texas, earned $30.6 million in power curtailment credits in Q3 2025 alone, a 147% increase year-over-year. In November 2025, a single month yielded $1 million in curtailment credits plus $1.3 million from formal demand response participation.

Why miners make ideal demand response participants: Mining load can curtail to near zero in seconds without equipment damage, lost product, or complex restart procedures. The CCAF reports that miners curtailed 888 GWh of electrical load during 2023, and 70.8% of surveyed miners actively undertake climate mitigation measures. No other industrial consumer offers this combination of scale and flexibility.

How demand response works for miners

The mechanism is straightforward. Miners sign contracts with grid operators or participate in ancillary services markets. When wholesale electricity prices spike (typically during extreme heat or cold), miners reduce or halt operations. They earn revenue in two ways: selling their contracted power back to the grid at spot prices, and receiving direct payments from demand response programs. During off-peak hours and periods of excess generation (especially from wind and solar), miners ramp back up, absorbing power that might otherwise be curtailed or wasted.

This creates a symbiotic dynamic. Renewable generators benefit from a buyer of last resort for excess production, improving project economics and enabling more renewable capacity to be built. Grid operators gain a reliable source of curtailable load. And miners access cheaper electricity by accepting interruptible supply. The CCAF study found this pattern most pronounced in regions with high renewable penetration, where generation variability creates frequent surplus periods.

ASIC Efficiency and the Hardware Arms Race

The efficiency of ASIC miners has improved dramatically. In 2018, a typical mining rig consumed approximately 100 joules per terahash (J/TH). Today's leading hardware, the Bitmain Antminer S23 series (presented at the World Digital Mining Summit in May 2025), achieves 9.5 J/TH with hydro cooling and 11.0 J/TH with air cooling. This represents roughly a 10x improvement in eight years.

The S23 produces 69% more hashrate and operates 41% more efficiently than the S21. However, the network-wide fleet average remains roughly 28 J/TH, reflecting the mix of old and new hardware deployed across operations. The difficulty adjustment mechanism ensures that as more efficient hardware comes online, the network adjusts, maintaining approximately 10-minute block times regardless of total hashrate.

Efficiency gains matter for the energy debate because they mean each terahash of network security requires less power. But the difficulty adjustment also means that efficiency savings tend to be reinvested as additional hashrate rather than reduced total consumption. This is sometimes called the Jevons paradox applied to mining: more efficient hardware leads to more mining, not less energy use.

The AI Pivot and Its Energy Implications

The most significant development in mining infrastructure during 2025-2026 is the industry's pivot toward AI and high-performance computing (HPC). Public mining companies have announced more than $70 billion in cumulative AI/HPC contracts. Core Scientific contracted roughly $10 billion through CoreWeave. IREN signed a $9.7 billion deal with Microsoft for 76,000 NVIDIA GB300 GPUs. Industry projections suggest listed miners could derive up to 70% of revenue from AI by the end of 2026.

This convergence has direct energy implications. Mining data centers are being repurposed or expanded for AI workloads, which require consistent power rather than the interruptible supply model that made miners useful for grid balancing. As mining companies transition to AI hosting, their value as flexible grid load may diminish, even as their total energy consumption increases. The tension between these two roles will shape energy policy debates around mining infrastructure going forward.

Addressing the Criticisms

Total energy consumption

At approximately 138 TWh per year, Bitcoin's energy consumption is substantial. Critics argue that this energy could be better allocated to other uses, regardless of whether it comes from renewable sources. Proponents counter that the comparison should be against the systems Bitcoin aims to supplement or replace, and that mining's flexibility as interruptible load adds value to grids beyond simple consumption. Both arguments have merit: the energy is real, but so is the grid-balancing service and the security it provides to a global settlement network.

E-waste from ASIC hardware

ASICs are single-purpose machines that cannot be repurposed for other computing tasks. A 2021 study by de Vries and Stoll, published in Resources, Conservation and Recycling, estimated that Bitcoin mining generated approximately 30.7 metric kilotons of e-waste annually, with an average economic lifespan of 1.29 years per unit before replacement. While these figures predate the current generation of more efficient ASICs (whose longer competitiveness at smaller process nodes may extend useful lifespans), e-waste remains a genuine environmental concern that the industry has not fully addressed.

Opportunity cost of energy

Even when miners use renewable energy, critics point to opportunity cost: that energy could power homes, hospitals, or other industries. This argument is strongest in regions with constrained grids (Norway's 2025 policy shift against new mining facilities reflects exactly this concern) and weakest where miners consume genuinely stranded or surplus energy that has no alternative buyer. The distinction between “renewable” and “additional renewable” matters: miners using existing hydro capacity are not adding new clean energy to the grid.

Comparison with traditional banking

A widely cited 2021 Galaxy Digital estimate placed the traditional banking system's energy consumption at approximately 263 TWh annually, including data centers, physical branches, ATMs, and card networks. However, this comparison has significant limitations. The banking figure is an industry estimate, not independently verified. More importantly, the traditional banking system serves billions of users and processes vastly more transactions. Bitcoin L1 and traditional banking serve fundamentally different functions: Bitcoin provides final settlement for a global asset network, while banking provides full-service retail/commercial financial services. Direct per-transaction comparisons are misleading in both directions.

Layer 2 Solutions Change the Energy-per-Payment Equation

Bitcoin's energy debate is often conflated with transaction efficiency. The network's energy consumption secures the base layer and produces new coinbase transactions through block subsidies, independent of how many transactions each block contains. Layer 2 solutions fundamentally change the economics: they enable millions of additional payments without requiring additional energy per transaction at the base layer.

Spark, for instance, enables instant, self-custodial Bitcoin and stablecoin transfers off-chain. Each transfer on Spark settles without broadcasting an on-chain transaction, meaning the energy cost per payment approaches zero beyond what the base layer already consumes for network security. The same base-layer energy budget that secures a few thousand L1 transactions per block can underpin millions of L2 payments. This reframes the energy debate: the relevant metric is not energy per L1 transaction but energy per unit of economic activity secured by the network.

For a deeper look at the post-halving economics driving miner behavior, see our analysis of Bitcoin mining economics in 2026. Developers interested in building on Layer 2 infrastructure can explore the Spark documentation and SDK.

What Comes Next

Several forces will shape Bitcoin mining's energy trajectory through 2026 and beyond. Nuclear power's share of the mining energy mix grew from roughly 4% in 2021 to nearly 10% in 2025, and interest continues to grow. The AI/HPC convergence may redirect mining infrastructure toward less flexible, higher-consumption computing workloads. The April 2024 halving has already pushed marginal miners offline, concentrating hashrate among operators with the lowest energy costs and most efficient hardware.

Policy is evolving unevenly. The United States passed the GENIUS Act and repealed restrictive accounting guidance (SAB 121) in favor of the more accommodating SAB 122, while the EU fully implemented its MiCA regulation. Countries like Ethiopia and Paraguay are tightening mining-specific regulations even as others (Oman, UAE) actively court mining operations. The next chapter of Bitcoin's energy story will be written not just by miners and hardware manufacturers, but by grid operators, regulators, and the competing demands of AI infrastructure.

This article is for educational purposes only. It does not constitute financial or investment advice. Bitcoin and Layer 2 protocols involve technical and financial risk. Always do your own research and understand the tradeoffs before using any protocol.