Bitcoin Energy Consumption Statistics 2026: Efficiency, Mix and TWh

Bitcoin’s network drew an estimated 138 terawatt-hours of electricity over the most recent year, roughly 0.5% of global consumption, according to the Cambridge Centre for Alternative Finance. The live Cambridge Bitcoin Electricity Consumption Index runs a hashrate-weighted hardware-basket model with Power Usage Effectiveness bounds of 1.01 at the floor and 1.20 at the ceiling, and the Digiconomist economic model lands higher at 204.44 TWh. These Bitcoin energy consumption statistics share one honest caveat: the answer is a range, not a single number, and the gap is wide enough to matter.

Key Takeaways

• The figure stands, per Cambridge, at 138 TWh, about 0.5% of global consumption.
• The Digiconomist index reads 204.44 TWh, a roughly 48% higher figure driven by a different economic methodology.
• Sustainable sources power 52.4% of Bitcoin mining, combining 42.6% renewables with 9.8% nuclear.
• The most efficient air-cooled ASIC in 2026 runs at 13.5 J/TH, down from 98 J/TH for the 2016-era S9.
• Per EIA estimates, US-based mining drew 0.6% to 2.3% of all US electricity in 2023, or 25 TWh to 91 TWh.
• Cambridge revised its own 2021 estimate down 14.4%, from 104.0 TWh to 89.0 TWh, after re-weighting mining hardware.
• The April 2024 halving cut block rewards from 6.25 BTC to 3.125 BTC, pressuring less efficient miners off the network.

Editor’s Choice

• Cambridge’s network-wide emissions estimate stands at 39.8 MtCO2e per year.
• The carbon footprint figure reaches, per Digiconomist, 114.03 Mt CO2, comparable to the Czech Republic.
• A single Bitcoin transaction carries an estimated energy footprint of 845.23 kWh under the Digiconomist model.
• The US hosts 75.4% of reported global mining activity, with Canada second at 7.1%.
• The Bitcoin network hashrate surged past 800 EH/s in 2026, with some readings approaching 1,000 EH/s.
• Natural gas now supplies 38.2% of mining energy, overtaking coal at 8.9%.
• Bitcoin’s annual electricity use is an estimated 138 TWh (Cambridge best estimate).

Recent Developments

• June 2026: The Digiconomist Bitcoin Energy Consumption Index read 204.44 TWh, comparable to Thailand’s national power draw.
• May 2026: Network hashrate readings climbed toward 1,000 EH/s, sustaining upward pressure on aggregate electricity demand.
• April 2026: The most efficient commercial ASICs reached 13.5 J/TH air-cooled, with hydro-cooled machines pushing toward 9 to 10 J/TH.
• April 2025: The Cambridge Digital Mining Industry Report put sustainable energy at 52.4% of the mining mix, up from 37.6% in 2022.
• April 2025: Cambridge reported network-wide emissions of 39.8 MtCO2e against 138 TWh of annual consumption.

How Much Energy Does Bitcoin Use Per Year?

According to Cambridge, Bitcoin’s annual electricity use sits between 138 TWh (its survey-based best estimate, about 0.5% of global consumption) and, per Digiconomist, 204.44 TWh, a level comparable to the power consumption of Thailand. The live Cambridge CBECI sits between those poles by design, with a best-guess track that applies a Power Usage Effectiveness of 1.10 to a weighted basket of profitable hardware, bounded below by PUE 1.01 and above by 1.20.

• Cambridge’s report frames consumption as 138 TWh, or about 0.5% of global electricity consumption.
• The methodology, per Cambridge, publishes three bounds, applying Power Usage Effectiveness of 1.01 at the lower bound and 1.20 at the upper bound.
• Digiconomist’s economic model reaches 204.44 TWh, comparable to the power consumption of Thailand.
• Cambridge’s network-wide emissions land at 39.8 MtCO2e per year.

The CBECI updates every 24 hours and applies a seven-day moving average to its annualized figures for stability. A reader looking for one clean number will not find one; the responsible framing is a band anchored to the methodology behind each endpoint.

Don't chase the news. Let us curate it.

You get one weekly briefing with only the stories that matter. If the market is quiet, we skip it.

Email address

Get the Weekly Briefing - it's free!

✅ Join readers from Visa, Vanguard, and the FDIC.

Why Bitcoin Energy Estimates Disagree

The roughly 48% gap between the lowest and highest estimate is why Bitcoin energy consumption statistics vary so widely: the difference comes from methodology, not measurement error, because no central meter records the global network’s draw. Each estimator infers consumption from different inputs.

Source: Cambridge Centre for Alternative Finance, Digiconomist

• The Cambridge best-guess model applies a weighted basket of profitable hardware models with a Power Usage Effectiveness of 1.10, weighted by hardware deployment dates over five-year depreciation periods.
• Cambridge assumes miners are rational economic agents who only operate their devices for as long as they are profitable.
• The Digiconomist model instead assumes miners spend a certain proportion of their revenues on electricity, producing a higher estimate than hashrate-derived models.
• The spread between Cambridge’s 138 TWh and Digiconomist’s 204.44 TWh is about 66 TWh.

Why it matters: Cambridge’s hashrate-weighted model and Digiconomist’s revenue-based model diverge by roughly 66 TWh on the same network. Cambridge applies a Power Usage Effectiveness of 1.10 to a weighted hardware basket; Digiconomist assumes miners spend a fixed share of revenue on power. Neither is wrong, but readers should cite the method alongside the number.

The same modeling sensitivity shows up across the broader cryptocurrencies sector, where on-chain activity is measurable but energy and cost inputs must be inferred.

Bitcoin Mining Energy Mix

Sustainable sources powered an estimated 52.4% of Bitcoin mining as of the April 2025 Cambridge report, combining 42.6% renewables with 9.8% nuclear. The mix has shifted markedly since 2022, when fossil fuels dominated.

• Sustainable energy reached 52.4% of the mix, up from 37.6% in 2022.
• Natural gas climbed to 38.2%, up from 25.0% in 2022, replacing coal as the single largest source.
• Coal fell to 8.9%, down sharply from 36.6% in 2022.
• The findings rest on a survey of 49 digital mining firms representing 48% of global mining activity.

Is Bitcoin bad for the environment?

Bitcoin mining carries a real carbon cost, but the picture is shifting. Cambridge estimates network-wide emissions at 39.8 MtCO2e per year against a 52.4% sustainable energy mix, while Digiconomist’s model places the footprint higher at 114.03 Mt CO2, comparable to the Czech Republic. The trend toward gas and renewables is lowering per-unit intensity even as total draw rises.

How Bitcoin’s Energy Use Compares to Countries

Bitcoin’s electricity draw lands in mid-sized-nation territory, though which nation depends on the estimate used. The comparison is a useful frame because it converts an abstract TWh figure into something readers can picture.

• The Digiconomist index pegs annual draw at 204.44 TWh, comparable to the power consumption of Thailand.
• Its carbon figure of 114.03 Mt CO2 is comparable to the carbon footprint of the Czech Republic.
• Cambridge’s estimate equals 138 TWh, or about 0.5% of global electricity consumption.
• Digiconomist also estimates fresh water consumption of 3,222 GL annually, comparable to the total water use of Switzerland.

Mining Efficiency Has Improved 85% in a Decade

Per-hash efficiency improved roughly 85% in a decade, from 98 J/TH for the 2016-era Antminer S9 to 13.5 J/TH for the most efficient air-cooled machine in 2026. Yet total network consumption rose, because hashrate grew faster than efficiency gains.

• The 2016 Antminer S9 ran at 98 J/TH, while the 2020 generation reached 30 to 40 J/TH.
• The 2026 Bitmain Antminer S21 XP runs at 13.5 J/TH and 270 TH/s, the most efficient air-cooled Bitcoin miner commercially available.
• The best hydro-cooled machines are pushing toward 9 to 10 J/TH.
• Machines above 25 J/TH now struggle at typical European electricity rates.

Key finding: A decade of hardware progress cut the energy needed per terahash by roughly 85%, from 98 J/TH on the 2016 Antminer S9 to 13.5 J/TH on the 2026 S21 XP. Aggregate consumption still rose because network hashrate climbed past 800 EH/s, outpacing the per-unit efficiency gains and absorbing them at the network level.

The network hashrate surged past 800 EH/s in 2026, with some readings approaching 1,000 EH/s, which is the central reason efficiency gains have not lowered total draw. This same hardware-upgrade pressure runs through the broader crypto mining economy.

Bitcoin Mining and the US Power Grid

US-based Bitcoin mining drew an estimated 0.6% to 2.3% of all United States electricity demand in 2023, which was 3,900 TWh. That range translates into a wide absolute band reflecting how concentrated and opaque the sector remains.

• The EIA estimated US mining at 25 TWh to 91 TWh annually.
• It identified a total of 137 facilities, with 52 mapped by location and capacity.
• These sites cluster in 21 states, with most in Texas, Georgia, and New York.
• Globally, Cambridge placed the US at 75.4% of reported mining activity, followed by Canada at 7.1%.

The EIA has discontinued the emergency collection of data for Form EIA-862, the Cryptocurrency Mining Operations Survey. That leaves the federal figures frozen at a 2023 snapshot, so the US share has likely shifted since, and any 2026 citation of these numbers should carry that caveat. The geographic concentration tracks broader crypto adoption rates by country, where the United States consistently ranks among the largest markets.

Per-Transaction Energy and Why It Misleads

A single Bitcoin transaction carries an estimated energy footprint of 845.23 kWh, equivalent to the power consumption of an average US household over 28.97 days under the Digiconomist model. The figure is technically derived but conceptually misleading.

• The per-transaction figure stands at 845.23 kWh.
• That equals the power use of an average US household over 28.97 days.

Bitcoin’s energy spend secures the network, not individual transactions; the same mining draw protects the chain whether it processes one transaction or a million in a block. Dividing total energy by transaction count produces a large, alarming number that tells you almost nothing about marginal cost. The security budget scales with hashrate and the 3.125 BTC block reward set by the April 2024 halving, not with throughput.

How many TWh does Bitcoin consume?

The leading Bitcoin energy consumption statistics for 2026 range from 138 TWh (Cambridge’s survey-based figure) to 204.44 TWh (Digiconomist), with the live Cambridge CBECI hashrate model publishing three bounds (PUE 1.01 floor, 1.10 best-guess, 1.20 ceiling) updated every 24 hours. The right answer is a range, and citing the methodology matters as much as the number.

What percentage of Bitcoin mining uses renewable energy?

Renewables such as hydro, wind, and solar supplied 42.6% of mining energy as of the April 2025 Cambridge report, rising to 52.4% sustainable when 9.8% nuclear is included. Natural gas, at 38.2%, is the largest single source.

Why do Bitcoin energy estimates differ so much?

Estimates differ because no meter records the network’s global draw, so each model infers it from different inputs. Digiconomist’s revenue-based economic model produces a higher estimate than hashrate-derived models like Cambridge’s, creating a spread of roughly 66 TWh between the two.

Conclusion

Bitcoin’s energy footprint in 2026 is best read as a band: 138 TWh, or about 0.5% of global electricity, at the Cambridge survey end, with the live CBECI hashrate-weighted model bracketing a wider range via its PUE 1.01 to 1.20 bounds, and 204.44 TWh at the Digiconomist end, against a sustainable energy mix that has climbed to 52.4%. The honest takeaway for investors and policy researchers tracking SEC and CFTC crypto regulation data is that the number depends on the method, and the most-cited “as much power as a country” headlines rarely say which country or which model.

The trajectory worth watching is the decoupling of efficiency from total draw. As ASICs approach 10 J/TH and the sustainable mix passes half, the per-unit story improves even as aggregate consumption tracks hashrate higher, a tension that will define the network’s environmental debate through the next halving cycle.