Skip to content
COLD HARDSATS
Get the brief

Learn · Energy

Will bitcoin boil the oceans?

Bitcoin's electricity use, what it compares to, where the power comes from, what critics worry about, and the scary predictions that missed. Every number is sourced and dated.

The short answer

No. On October 17, 2026, the Cambridge Bitcoin Electricity Consumption Index (CBECI) put bitcoin's electricity use at about 169 TWh a year, a steady draw of about 19 GW.[1] That's about 0.5% of the electricity the world used in 2025 (about 31,734 TWh), or roughly 1 kWh in every 190.[15] A lot, but not planet-scale.

Perspective · the ocean math

  • • The oceans hold about 1.335 billion cubic kilometers of water,[2] at least 1.3 × 1021 kg.
  • • Warming seawater by 1 °C takes about 3,990 joules per kilogram.[3] For all the oceans, that's about 5.3 × 1024 J, or about 1.5 billion TWh.
  • • At 169 TWh a year, bitcoin would need about 8.8 million years of electricity to warm the oceans by just 1 °C, even if every bit of it ended up as ocean heat and none escaped to space. Boiling them would take far more.
  • • Climate change is already adding heat on a vastly bigger scale: the upper 2,000 m of the oceans gained about 16 ± 8 zettajoules (1021 J) in 2024.[4] Bitcoin's whole year of electricity is about 0.0006 ZJ, so more than 10,000 times smaller, even at the low end of that range.

Our arithmetic from the cited inputs, using fresh-water density as a floor (seawater is slightly denser).

The direct heat is a red herring. The real climate question is emissions: Cambridge estimates bitcoin emitted about 95.5 MtCO2e in 2025,[5] about 0.17% of the 57.7 GtCO2e of greenhouse gases the world emitted in 2024.[6] Whether that's too much is a values question, and both sides are below.

How much electricity does bitcoin use?

No one can meter it directly. Miners are spread around the world and many don't report. Every estimate is a model, and the models disagree. Cambridge starts from the network's computing power (hashrate) and assumes miners only run machines that are profitable, then works out the power that mix of real hardware would draw.[7] Its range is wide because the true mix of machines is unknown: the low end assumes only the most efficient machines, the high end only the least efficient.[7] Digiconomist, run by long-time critic Alex de Vries, instead works backwards from what miners earn and what they likely spend on electricity.[9][43]

Estimates of bitcoin's electricity use
EstimateTWh / yearDateMethod
Cambridge CBECI best guess[1]168.9 (86.6–324.9)Oct 17, 2026Hardware model; live annualized rate
Cambridge CBECI, full year 2025[1]188.62025Sum of the daily estimates
Cambridge miner survey[11][12]138Apr 2025 reportSurvey of 49 firms covering 48% of mining
Digiconomist (critics' estimate)[9][10]204.4Oct 2026Economic model from miner revenue
US mining only, EIA[13]25–912023Preliminary; 0.6–2.3% of US use

The spread is real: Digiconomist's figure is about 21% above Cambridge's live best guess, and Cambridge's own range runs from about half to almost double its central number.[1][9] The Cambridge index peaked at about 227 TWh (annualized) in October 2025.[1] The US government tried to measure it directly. In 2024 the EIA started an emergency survey of US miners, but after a lawsuit from the Texas Blockchain Council and Riot Platforms, it agreed to destroy the data it had collected and propose a new survey.[14]

Compared with everyday things

Cambridge warns that every comparison has three limits: apples and oranges, limited data, and "presenter bias." The same number can look tiny or huge depending on what you put next to it.[16] So here are many comparisons from both directions, each dated and sourced.

Electricity per year, in terawatt-hours (TWh)

Orange bars are bitcoin estimates; the faint band shows each estimate's range. Years differ, so treat this as orders of magnitude, not a ranking. 1 TWh = 1 billion kWh.

  • All data centres worldwide (2024)[17]
    415
  • Banking system (Galaxy estimate) (2021 est.)[18]
    264
  • Gold industry (Galaxy estimate) (2021 est.)[18]
    241
  • Bitcoin: Digiconomist (Oct 2026)[9][10]
    204
  • Poland (2025)[15]
    174
  • Bitcoin: Cambridge (CBECI) best guess (Oct 17, 2026)[1][7]
    169
  • Argentina (2025)[15]
    162
  • Always-on idle devices in US homes (2015 est.)[19]
    ≈150
  • Bitcoin: Cambridge miner survey (2025 report)[11]
    138
  • Gold mining (Cambridge, from Mudd) (2019 est.)[16]
    131
  • Netherlands (2025)[15]
    121
  • Bitcoin mining in the US (EIA) (2023)[13]
    25–91
  • Clothes dryers in US homes (2020)[20]
    56
  • US holiday lights (2007 est.)[21]
    6.6
  • Netflix’s own operations (2023)[22]
    0.11
Not shown to scale: the whole world used about 31,734 TWh of electricity in 2025 (Ember), a bar about 75 times wider than data centres' and about 190 times wider than Cambridge's bitcoin estimate.[15] The faint band on Cambridge's bar is its range (87–325 TWh); the US mining figure is a range only (25–91 TWh).

Perspective · using Cambridge's 169 TWh

  • • About 25 times the electricity used by all US decorative holiday lights, using the Department of Energy's 2007 estimate of 6.63 TWh, made when only 5.2% of those lights were LEDs.[21]
  • • About 3 times what all US home clothes dryers use (about 56 TWh in 2020).[20]
  • • About the same as the electricity US homes waste on always-on devices nobody is using (NRDC's 2015 estimate, about $19 billion a year).[19]
  • • About 40% of what the world's data centres used in 2024 (415 TWh), and they're growing fast.[17]
  • • Between Argentina (162 TWh) and Poland (174 TWh) in 2025.[15]

Streaming and Netflix. Netflix reported using 113,753 MWh (about 0.11 TWh) in 2023, but that covers only its own servers, offices and productions.[22] The IEA found that most of the energy of streaming is used by viewers' own screens (about 72%), then networks (23%) and data centres (5%). It estimated a typical hour of Netflix in 2019 used about 0.077 kWh.[23] The IEA also showed how a widely shared 2019 claim, that 30 minutes of Netflix emits as much as driving almost 4 miles, overstated the impact by up to 90 times.[23] We found no sourced, current TWh total for all streaming worldwide, so it isn't in the chart.

Online adult video. In 2019 the French think tank the Shift Project estimated that pornography made up 27% of online video traffic and produced more than 80 MtCO2 in 2018, "as much as all France's households."[24] It gave that figure as emissions, not electricity, so it isn't in the TWh chart. For comparison, Cambridge estimates bitcoin emitted about 23 MtCO2e in 2018.[5] Treat the Shift Project number with care. The IEA found its model overstated data-centre energy about 35-fold and network energy about 50-fold, while understating viewers' devices about 4-fold.[23] The Shift Project says the bit/byte error it corrected in 2020 came from an interview figure, not its published report.[25]

Gold and banks. These are the comparisons bitcoin fans like best, and the shakiest. Galaxy Digital, a firm with a bitcoin-mining business, estimated the banking system at 263.72 TWh and gold at 240.61 TWh in 2021. It said bank data is "scarce" and that comparing bitcoin with the layers of the banking system is "inherently flawed."[18] Its gold figure was converted from emissions using a world-average carbon intensity.[18] Cambridge's own gold estimate is lower, 131 TWh, and rests on 2006 data that it says may be out of date.[16] One more caution: the same Galaxy report lists 1,375 TWh as US "always-on" device use. In NRDC's report, 1,375 billion kWh is the total for all electricity used in US homes.[18][19] Always check the original source.

Where the power comes from

For the climate, where the power comes from matters as much as how much is used. The estimates again depend on who's counting:

  • Cambridge's 2025 miner survey: 52.4% sustainable (42.6% renewables, 9.8% nuclear), up from Cambridge's 2022 estimate of 37.6%. Natural gas, at 38.2%, has replaced coal (8.9%) as the largest single source.[11]
  • Bitcoin Mining Council (an industry group): its members reported a 63.1% sustainable mix for the first half of 2023, and it estimated 59.9% for the whole network. The figures are self-reported by member companies. That's the latest council survey we could verify; its website didn't load when we checked on October 17, 2026.[26]

Flare gas. Oil wells far from pipelines often burn off the natural gas that comes up with the oil. The World Bank counted 151 billion cubic meters flared worldwide in 2024, the most since 2007.[27] Mining rigs can sit at the well and turn some of that gas into electricity.

Stranded and curtailed power. Miners can move and can switch off when power is scarce, so they can use power that would otherwise go to waste: in the past, surplus wet-season hydro in China, and today, gas that would be flared.[16] In Cambridge's survey, miners reported cutting 888 GWh of load in 2023 when grids needed it.[12]

Grid demand response. In Texas, big flexible users can be paid to switch off when the grid is strained. Bitcoin miners are one such user, along with steel mills, battery operators and others.[29]

Methane. Methane is a far stronger greenhouse gas than CO2 over the short term, so burning landfill or oil-field gas that would otherwise leak can cut overall warming. That's the case pilots like the one below are testing.[30] In 2023 Cambridge said flare-gas and landfill-methane mining could lower its emissions estimates, which didn't yet account for them.[34]

The critics' case

  • Emissions are real and rising. Cambridge's yearly emissions estimate rose from about 33 MtCO2e in 2020 to about 96 MtCO2e in 2025.[5] Digiconomist's current estimate is higher, about 114 MtCO2 a year.[9] Even Cambridge's survey found 47.6% of surveyed miners' electricity came from fossil fuels.[12]
  • E-waste. Specialized mining machines become obsolete as newer, more efficient ones arrive. De Vries and Stoll estimated 30.7 thousand tonnes of e-waste a year as of May 2021, about 272 g per transaction, assuming machines last about 1.29 years.[33] Digiconomist's current estimate is 23.12 thousand tonnes.[9] Lifespans are disputed: Cambridge notes that academic estimates run around 1.5 years or less, while industry estimates are usually 3 to 5 years.[34]
  • Noise and local impacts. Big mines run cooling fans day and night. Cambridge also lists noise, water use and thermal pollution as impacts its emissions estimates don't capture.[34]
  • Grid strain. The EIA said electricity demand from US crypto mining "has grown very rapidly." Its low-end estimate equals the yearly electricity use of whole states like Utah or West Virginia.[13]
  • Flare-gas mining is still fossil fuel. Critics argue it can turn burning gas into a new business for drillers rather than a climate fix.[31]

The "per transaction" debate

In 2018, de Vries wrote that a single bitcoin transaction used as much electricity as an average Dutch household does in a month.[43] Digiconomist currently puts it at 773.62 kWh per transaction.[9] The number is real arithmetic, total energy divided by transactions, but what it means is disputed.

Why Cambridge calls it misleading: "Bitcoin's energy footprint is linked to block production, not transaction processing." A full block and an empty block cost the same energy at a given difficulty. One transaction can also bundle hundreds of payments or settle thousands of off-chain Lightning payments.[8] Adding more transactions doesn't add energy, and cutting them wouldn't save any.

The critics' reply: Digiconomist argues that the network's transaction capacity is capped, about 220 million transactions a year, while its energy use isn't, because price drives it. It also argues that moving funds into Lightning still needs an on-chain transaction, so bitcoin can't scale to everyday payments without trusted third parties.[9] The fair summary: per-transaction figures are poor for comparing payment systems, but they do show how much energy the system spends compared with how much it's used on-chain.

Predictions that didn't come true

Forecasting bitcoin's energy use is hard. Cambridge says a credible forecast would first need an accurate prediction of bitcoin's price.[8] Here's how some well-known forecasts held up against the data.

  1. 1.Did not happen

    ClaimIf the growth rate of late 2017 continued, bitcoin would use “as much energy as the U.S. by 2019, and as much energy as the entire world by the end of 2020.”[36]

    Who and whenNewsweek, Dec 11, 2017, extrapolating Digiconomist's index. The article itself called the projection “purely hypothetical” and noted a similar ZeroHedge estimate of February 2020.

    What happenedCambridge estimates bitcoin used about 55 TWh in 2019, about 1.3% of US demand that year (about 4,197 TWh), and about 67 TWh in 2020, about 0.25% of the world's electricity (about 26,713 TWh).[1][15]

  2. 2.Far off

    Claim“Projected Bitcoin usage, should it follow the rate of adoption of other broadly adopted technologies, could alone produce enough CO2 emissions to push warming above 2 °C within less than three decades.”[37]

    Who and whenCamilo Mora and colleagues at the University of Hawai‘i, in Nature Climate Change, Oct 29, 2018. Two rebuttals ran in the same journal on Aug 28, 2019, and the authors replied the same day.[38][39][40]

    What happenedDittmar and Praktiknjo calculated that the median scenario (about 231.4 GtC within 16 years) would require tripling world electricity generation within five years. Masanet and five co-authors found the model used outdated hardware and grid data; fixing those cut its 2017 starting point from 69 to about 15.7 MtCO2e. Cambridge now estimates bitcoin's 2017 emissions at about 6.9 MtCO2e and its total emissions from 2010 through 2025 at about 432 MtCO2e. That's about 0.12 GtC, or roughly 1/2,000 of the scenario's 16-year total, halfway through that window.[38][39][41][5]

  3. 3.Roughly right, early

    ClaimBitcoin used at least 2.55 GW and “could be using as much as 7.7 gigawatts” by the end of 2018, about half a percent of world electricity. He added that it could someday reach 5%.[42]

    Who and whenAlex de Vries (Digiconomist), Joule, May 16, 2018.[43]

    What happenedCambridge's best guess averaged about 3.9 GW in December 2018 and first reached 7.7 GW on Sep 3, 2019, nine months late. Bitcoin's share of world electricity reached about 0.5% in 2024 (159 of about 30,913 TWh), six years later than predicted. Today's estimate is 19.3 GW. The level was right; the timing was early. The 5% scenario hasn't happened; the share is about 0.6%.[1][15]

  4. 4.Revised down

    ClaimEstimates move too: Cambridge's own model once put 2021 at 104.0 TWh.[34]

    Who and whenCambridge Centre for Alternative Finance, which revised its method in August 2023.

    What happenedAfter checking import records and hardware sales data, Cambridge lowered 2021 to 89.0 TWh and 2022 from 105.3 to 95.5 TWh. It said its old method overweighted older machines when mining was very profitable. Honest estimates get revised, so treat every figure on this page as a best estimate, not a measurement.[34]

Being fair cuts both ways. De Vries's level was roughly right, just years early, and bitcoin's electricity use has kept rising: Cambridge's yearly total roughly quadrupled from 2018 (43 TWh) to 2025 (189 TWh).[1] The missed predictions don't mean bitcoin's footprint is small; they mean straight-line extrapolation is a poor guide.

So, will it boil the oceans?

No. Bitcoin uses about as much electricity as a mid-sized country, roughly 0.5% of the world's, and produces something like 0.07% to 0.2% of global greenhouse gas emissions, depending on whose estimate you use.[1][11][9][6] That's a meaningful footprint worth measuring honestly, not a planet-ending one. Whether it's worth it depends on how much you value what that energy buys: a ledger that's expensive to rewrite. For how that security works, see proof-of-work and mining.

Sources

  1. [1] Cambridge Bitcoin Electricity Consumption Index (CBECI), CCAF, live estimate read Oct 17, 2026
  2. [2] NOAA, “How much water is in the ocean?”
  3. [3] TEOS-10, gsw_cp0 (reference specific heat of seawater, 3991.87 J/kg/K)
  4. [4] L. Cheng et al., “Record High Temperatures in the Ocean in 2024”, Advances in Atmospheric Sciences (2025)
  5. [5] CBECI greenhouse gas emissions estimates (yearly totals, power mix), CCAF
  6. [6] UNEP, Emissions Gap Report 2025 (global GHG emissions 2024)
  7. [7] CBECI methodology, CCAF
  8. [8] CBNSI FAQ: “Does Bitcoin’s energy consumption rise proportionally with the number of transactions?”, CCAF
  9. [9] Digiconomist, Bitcoin Energy Consumption Index (read Oct 17, 2026)
  10. [10] Digiconomist API documentation (daily network kWh by date)
  11. [11] Cambridge Judge Business School, “Cambridge study: sustainable energy rising in Bitcoin mining” (Apr 28, 2025)
  12. [12] Cambridge Digital Mining Industry Report (CCAF, April 2025, PDF)
  13. [13] US EIA, “Tracking electricity consumption from U.S. cryptocurrency mining operations” (Feb 1, 2024)
  14. [14] American Public Power Association, “EIA Enters Agreement Tied to Collection of Cryptocurrency Mining Data” (2024)
  15. [15] Ember, Yearly Electricity Data (electricity demand by country and world, 2025 release)
  16. [16] CBECI comparisons page (comparison limits; gold, US appliance and country figures), CCAF
  17. [17] IEA, Energy and AI, “Energy demand from AI” (2025)
  18. [18] Galaxy Digital Mining, “On Bitcoin’s Energy Consumption: A Quantitative Approach to a Subjective Question” (May 2021, PDF)
  19. [19] NRDC, “Home Idle Load: Devices Wasting Huge Amounts of Electricity When Not in Active Use” (May 2015, PDF)
  20. [20] US EIA, 2020 Residential Energy Consumption Survey, Table CE5.1a (household electricity end uses, PDF)
  21. [21] US DOE (Navigant Consulting), “Energy Savings Estimates of Light Emitting Diodes in Niche Lighting Applications” (2008, PDF; 2007 data)
  22. [22] Netflix, 2023 Environmental Social Governance Report (PDF, electricity summary table)
  23. [23] G. Kamiya, IEA, “The carbon footprint of streaming video: fact-checking the headlines” (updated Dec 11, 2020)
  24. [24] The Shift Project, “Climate crisis: the unsustainable use of online video”, executive summary (July 2019, PDF)
  25. [25] The Shift Project, “Did The Shift Project really overestimate the carbon footprint of online video?” (June 2020)
  26. [26] Bitcoin Mining Council, H1 2023 survey release (Aug 2023; site unreachable when checked Oct 17, 2026)
  27. [27] World Bank, Global Gas Flaring Tracker Report (July 2025)
  28. [28] Riot Platforms, “Riot Announces August 2023 Production and Operations Updates” (Sep 6, 2023)
  29. [29] Riot Platforms, “Riot Responds to Recent Inquiries Regarding Its Power Strategy” (Sep 8, 2023)
  30. [30] CNBC, “Exxon is mining bitcoin in North Dakota as part of its plan to slash emissions” (Mar 26, 2022)
  31. [31] DeSmog, “As Oil Giants Turn to Bitcoin Mining, Some Spin Burning Fossil Fuels for Cryptocurrency as a Climate Solution” (Mar 31, 2022)
  32. [32] Nodal Power, “Nodal Power X MARA LFG Learnings Report” (June 12, 2024)
  33. [33] A. de Vries & C. Stoll, “Bitcoin’s growing e-waste problem”, Resources, Conservation & Recycling 175 (2021)
  34. [34] A. Neumueller, “Bitcoin electricity consumption: an improved assessment”, Cambridge Judge Business School (Aug 31, 2023)
  35. [35] Earthjustice, “Granbury Residents Sue Local Bitcoin Mine Over Health-Threatening Noise Pollution” (Oct 2024)
  36. [36] A. Cuthbertson, “Bitcoin Mining on Track to Consume All of the World’s Energy by 2020”, Newsweek (Dec 11, 2017)
  37. [37] C. Mora et al., “Bitcoin emissions alone could push global warming above 2°C”, Nature Climate Change 8, 931–933 (Oct 29, 2018)
  38. [38] L. Dittmar & A. Praktiknjo, “Could Bitcoin emissions push global warming above 2 °C?”, Nature Climate Change (Aug 28, 2019)
  39. [39] E. Masanet et al., “Implausible projections overestimate near-term Bitcoin CO2 emissions”, Nature Climate Change (Aug 28, 2019)
  40. [40] Mora et al. reply, Nature Climate Change (Aug 28, 2019)
  41. [41] Masanet et al. 2019, free author version with supplementary analysis (Lawrence Berkeley National Laboratory, PDF)
  42. [42] A. de Vries, “Bitcoin’s Growing Energy Problem”, Joule 2(5), 801–805 (May 16, 2018)
  43. [43] Cell Press via ScienceDaily, “Bitcoin estimated to use half a percent of the world's electric energy by end of 2018” (May 16, 2018)
  44. [44] FOX 4 Dallas-Fort Worth, “Residents near Granbury file lawsuit against Bitcoin mining company” (Oct 2024)

Data credit: bitcoin electricity and emissions estimates are from the Cambridge Centre for Alternative Finance (CBECI, CC BY-NC-SA 4.0) and Digiconomist; country and world electricity from Ember. Cold Hard Sats doesn't produce these estimates; ratios and conversions on this page are our arithmetic from the cited figures.

Checked Oct 17, 2026 · Proof-of-work · FUD: energy