Notably: the article text, and referenced source (US Energy Information Administration), contradict the headline's claim. In the text, we see the "over 2%" (sp

Notably: the article text, and referenced source (US Energy Information Administration), contradict the headline's claim. In the text, we see the "over 2%" (specifically "2.3%") number is based on a survey and the conjecture "if [those facilities were] running all-out". But they don't run all-out! The article itself mentions facilities that shut down when grids under peak load. But also, if you click through to the linked US EIA study the article reports about – https://www.eia.gov/todayinenergy/detail.php?id=61364 https://www.eia.gov/todayinenergy/detail.php?id=61364 – you get the following verbatim undistorted information: "Our preliminary estimates suggest that annual electricity use from cryptocurrency mining probably represents from 0.6% to 2.3% of U.S. electricity consumption." – that is: not necessarily or truly "over 2%" as implied by ArsTechnica's deceptive headline & spin. Perhaps as low as 0.6%! "In addition, cryptocurrency miners in areas with fluctuating power prices have reduced their electricity use in response to periods of high prices in wholesale power markets, given the sensitivity of their operational profitability to electricity prices." "In practice, cryptocurrency mining facilities frequently run at less than their maximum designed capability." Bitcoin mining is specifically drawn to just those times & places where electricity is especially cheap, and stops on a dime the moment rates surge over what's mining-profitable, due to other peaking demand or sagging supply. Thus the strained assumption of always-running-full that's required for the "over 2%" headline is false. The EIA reports they "lack the data to provide a well-sourced estimate" of actual utilization, but at one point hand-waves a guess "80%". I suspect it's even lower, but even at EIA guess: 80% of 2.3% is less than, not "over", 2%. Ultimately, though any negative externalities of Bitcoin's energy consumption can and should be addressed by adding our best estimates of the cost of such externalities into the electric rates paid by miners – & all other electricity consumers. That solves not just Bitcoin's externalities – but those of every other energy user. Further, the way proof-of-work mining works, & difficulty adjusts, forces only those with the lowest-cost electricity to be profitable, in the long run. Difficulty rises to erase the profitability of everyone else. And in any sane energy system, the lowest-cost electricity will tend to be from sagging demand at mismatched-generation places & times: excess wind/nuclear/hydro power at night, excess solar power at peak daylight. That is: electricity that isn't (& for the moment couldn't) be used for other things, and is even sometimes sold at negative prices when grid/generation stability requires someone to take the load. In an ideal & rational generation grid, these regional/diurnal cycles could mean tons of electricity – far more than 2% of annual usage, by raw KWh, but not price paid – could be soaked up with countercyclical users like Bitcoin mining. And in so helping with utilities' fixed-costs & demand-smoothing, that consumption would primarily benefit other grid users' costs and reliability. It also provides de facto subsidies to cleaner-but-spikier generation (wind/solar) or cleaner-but-harder-to-adjust generation (nuclear/hydro) – relative to traditional fossil-fuels. That is, no surprisingly high "X%" (X < 50) of some superficially-disfavored electricity use should necessarily cause concern – it depends entirely on exactly when/where/how that X% is taken. Some uses, & generation, are "essentially free (or better)" with regard to any limited resources or externalities of concern.