{"course":{"slug":"ai-and-you","title":"AI, Safety and What Goes Wrong","url":"https://addaly.com/learn/ai-and-you","level":"beginner","track":"start"},"lesson":{"slug":"where-the-data-centre-lands","title":"Where the data centre lands","url":"https://addaly.com/learn/ai-and-you/where-the-data-centre-lands","position":66,"of":73,"minutes":9,"answer":"AI's environmental cost is concentrated and local (a fifth of Ireland's electricity, water consumed in drought regions, grid upgrades charged to household bills) and efficiency gains coexist with rising totals because cheaper computation increases demand.","keyIdea":"AI's environmental cost is concentrated and local (a fifth of Ireland's electricity, water consumed in drought regions, grid upgrades charged to household bills) and efficiency gains coexist with rising totals because cheaper computation increases demand."},"body":{"format":"markdown","text":"## The question is local, not global\n\nThe environmental conversation about AI usually happens at two useless altitudes: the individual query, which is negligible, and global emissions, which are too aggregated to act on. The decisions that matter sit in between, and they are decisions about where a building goes.\n\nThe International Energy Agency estimated global data centre electricity consumption at around 415 terawatt-hours in 2024, roughly 1.5% of world electricity, and projected it roughly doubling to about 945 terawatt-hours by 2030, close to Japan's total consumption today, with AI the fastest-growing component. Those are estimates with wide uncertainty bands, and the direction is not in doubt.\n\nBut a national or global percentage hides the thing that actually affects people, which is that this load is extremely concentrated. Data centres cluster where land, power, fibre and tax incentives coincide, and a handful of regions carry a share out of all proportion to their size.\n\n## What concentration does\n\n**Ireland** is the clearest case. Its Central Statistics Office reported data centres consuming around a fifth of all metered electricity in the country by 2023: a larger share than all urban households combined. The grid operator effectively stopped new connections in the Dublin region until the late 2020s. A national climate target now has to be met around a load that grew faster than the plan.\n\n> Figure: Ireland, 2023: who uses the metered electricity. Bars (% of metered electricity): Data centres 21; All urban households 18; All rural households 10. Data centres consumed a larger share than every urban household in the country combined, and the grid operator effectively stopped new connections in the Dublin region. A national percentage hides this; the load is concentrated where land, power, fibre and tax incentives coincide.\n\n**Northern Virginia** hosts the largest concentration in the world, and the utility's forecast load growth has become a public planning fight about new transmission lines and who pays for them.\n\n**Water** is the second constraint and it is more local still. Evaporative cooling consumes water (it does not merely borrow it), and the consumption happens where the building is. Google reported data centre water consumption in the billions of gallons annually, and Microsoft reported a sharp rise in the year large model training scaled up. Several facilities sit in water-stressed regions, including in Chile, Arizona and Spain, where local opposition has focused precisely on this. Some operators have moved to air cooling or closed-loop systems, which use more electricity, which is the trade.\n\n**Who pays for the grid.** When a large load arrives, transmission has to be built. Whether that cost falls on the company or is spread across every household bill is a regulatory decision, made in a proceeding almost nobody attends, and it is the most consequential AI policy question in many places.\n\n**Local generation.** Where grid connections are slow, operators have installed their own generation, sometimes gas turbines, sometimes before permits are finalised. A 2025 dispute in Memphis over turbines at a data centre in a neighbourhood already carrying poor air quality is the template for arguments that will recur.\n\n## Efficiency, and why it does not settle it\n\nData centres have become dramatically more efficient. Power usage effectiveness (total facility power divided by computing power) has fallen from around 2.0 two decades ago to about 1.1 in the best modern facilities, meaning overhead is now a tenth rather than a doubling. Chips do far more computation per watt each generation.\n\nAnd total consumption keeps rising, because efficiency lowers the cost of computation and lower cost raises demand. This is **Jevons paradox**, observed by William Stanley Jevons in 1865 about coal, and it is why per-unit efficiency claims and total consumption figures can both be honest and point in opposite directions. When a company reports improved efficiency, the useful follow-up is what happened to total consumption.\n\n## What is worth arguing about\n\nIf you want to engage with this beyond feeling bad about your own usage, the questions with leverage are these.\n\n**Is the electricity additional and clean?** A facility powered by newly built renewable capacity is a different proposition from one that buys certificates while drawing from a coal-heavy grid. Ask about hourly matching rather than annual offsetting, because a grid at 3am is not the grid at noon.\n\n**Who bears the grid cost?** Read the utility filing.\n\n**What is the water source and what is the local stress level?** Consumption in a water-abundant region is a different fact from the same consumption in a drought region.\n\n**Is there a community benefit that is not a jobs number?** Large data centres employ few people once built. Tax abatements are often granted against employment figures that never arrive.\n\nThe place your influence exists is as a citizen in a planning process, not as a user counting prompts."},"check":{"question":"An operator reports that its data centres have improved power usage effectiveness from 1.4 to 1.1 and describes this as a major environmental achievement. What is the necessary follow-up question?","options":[{"text":"Whether the measurement was independently verified, as self-reports are unreliable","correct":false,"tempting":"Verification is reasonable, but the figure could be accurate and still say nothing about total impact."},{"text":"Whether the gain came from better cooling or from newer chips, which last different times","correct":false,"tempting":"An interesting engineering detail that does not bear on whether the environmental claim means anything."},{"text":"Whether the sites run on renewable energy, which matters far more than efficiency","correct":false,"tempting":"An important question about carbon, but it leaves the efficiency claim unexamined and misses the demand effect."},{"text":"What happened to total electricity and water use, since cheaper compute draws demand","correct":true}],"why":"Efficiency lowers the cost of each unit of work, and lower cost usually raises the total. Only the totals answer the claim."},"prev":{"slug":"open-weights-and-accountability","title":"Open weights, and what open buys you","url":"https://addaly.com/learn/ai-and-you/open-weights-and-accountability"},"next":{"slug":"carbon-you-can-influence","title":"The choices that actually change the number","url":"https://addaly.com/learn/ai-and-you/carbon-you-can-influence"},"authorship":"Founder and series editor: Dr Mustafa Mun. Written by AI models to a written standard. Source and assessment review continues. Academic review invited.","review":{"status":"Draft, open for academic review"},"cite":"Addaly (2026) 'Where the data centre lands', in AI, Safety and What Goes Wrong. Edited by M. Mun. Addaly. Available at: https://addaly.com/learn/ai-and-you/where-the-data-centre-lands (Accessed: 21 November 2026).","licence":{"name":"CC BY-NC-SA 4.0","url":"https://creativecommons.org/licenses/by-nc-sa/4.0/"},"reader":{"schema":1,"title":"Where the data centre lands","courseTitle":"AI, Safety and What Goes Wrong","module":{"index":9,"slug":"living-with-it","title":"Living with it: cost, governance and your own practice"},"answer":null,"keyIdea":"AI’s environmental cost is concentrated and local (a fifth of Ireland’s electricity, water consumed in drought regions, grid upgrades charged to household bills) and efficiency gains coexist with rising totals because cheaper computation increases demand.","hero":{"kind":"figure","figure":{"t":"figure","label":"Figure 9.1","figure":{"caption":"Data centres consumed a larger share than every urban household in the country combined, and the grid operator effectively stopped new connections in the Dublin region. A national percentage hides this; the load is concentrated where land, power, fibre and tax incentives coincide.","series":[{"highlight":true,"label":"Data centres","value":21},{"label":"All urban households","value":18},{"label":"All rural households","value":10}],"title":"Ireland, 2023: who uses the metered electricity","type":"bars","unit":"% of metered electricity"},"alt":"Figure: Ireland, 2023: who uses the metered electricity. Bars (% of metered electricity): Data centres 21; All urban households 18; All rural households 10. Data centres consumed a larger share than every urban household in the country combined, and the grid operator effectively stopped new connections in the Dublin region. A national percentage hides this; the load is concentrated where land, power, fibre and tax incentives coincide."}},"blocks":[{"t":"h2","id":"the-question-is-local-not-global","runs":[{"text":"The question is local, not global"}]},{"t":"p","runs":[{"text":"The environmental conversation about AI usually happens at two useless altitudes: the individual query, which is negligible, and global emissions, which are too aggregated to act on. The decisions that matter sit in between, and they are decisions about where a building goes."}]},{"t":"p","runs":[{"text":"The International Energy Agency estimated global data centre electricity consumption at around 415 terawatt-hours in 2024, roughly 1.5% of world electricity, and projected it roughly doubling to about 945 terawatt-hours by 2030, close to Japan’s total consumption today, with AI the fastest-growing component. Those are estimates with wide uncertainty bands, and the direction is not in doubt."}]},{"t":"p","runs":[{"text":"But a national or global percentage hides the thing that actually affects people, which is that this load is extremely concentrated. Data centres cluster where land, power, fibre and tax incentives coincide, and a handful of regions carry a share out of all proportion to their size."}]},{"t":"h2","id":"what-concentration-does","runs":[{"text":"What concentration does"}]},{"t":"p","runs":[{"text":"Ireland","b":true},{"text":" is the clearest case. Its Central Statistics Office reported data centres consuming around a fifth of all metered electricity in the country by 2023: a larger share than all urban households combined. The grid operator effectively stopped new connections in the Dublin region until the late 2020s. A national climate target now has to be met around a load that grew faster than the plan."}]},{"t":"p","runs":[{"text":"Northern Virginia","b":true},{"text":" hosts the largest concentration in the world, and the utility’s forecast load growth has become a public planning fight about new transmission lines and who pays for them."}]},{"t":"p","runs":[{"text":"Water","b":true},{"text":" is the second constraint and it is more local still. Evaporative cooling consumes water (it does not merely borrow it), and the consumption happens where the building is. Google reported data centre water consumption in the billions of gallons annually, and Microsoft reported a sharp rise in the year large model training scaled up. Several facilities sit in water-stressed regions, including in Chile, Arizona and Spain, where local opposition has focused precisely on this. Some operators have moved to air cooling or closed-loop systems, which use more electricity, which is the trade."}]},{"t":"p","runs":[{"text":"Who pays for the grid.","b":true},{"text":" When a large load arrives, transmission has to be built. Whether that cost falls on the company or is spread across every household bill is a regulatory decision, made in a proceeding almost nobody attends, and it is the most consequential AI policy question in many places."}]},{"t":"p","runs":[{"text":"Local generation.","b":true},{"text":" Where grid connections are slow, operators have installed their own generation, sometimes gas turbines, sometimes before permits are finalised. A 2025 dispute in Memphis over turbines at a data centre in a neighbourhood already carrying poor air quality is the template for arguments that will recur."}]},{"t":"h2","id":"efficiency-and-why-it-does-not-settle-it","runs":[{"text":"Efficiency, and why it does not settle it"}]},{"t":"p","runs":[{"text":"Data centres have become dramatically more efficient. Power usage effectiveness (total facility power divided by computing power) has fallen from around 2.0 two decades ago to about 1.1 in the best modern facilities, meaning overhead is now a tenth rather than a doubling. Chips do far more computation per watt each generation."}]},{"t":"p","runs":[{"text":"And total consumption keeps rising, because efficiency lowers the cost of computation and lower cost raises demand. This is "},{"text":"Jevons paradox","b":true},{"text":", observed by William Stanley Jevons in 1865 about coal, and it is why per-unit efficiency claims and total consumption figures can both be honest and point in opposite directions. When a company reports improved efficiency, the useful follow-up is what happened to total consumption."}]},{"t":"h2","id":"what-is-worth-arguing-about","runs":[{"text":"What is worth arguing about"}]},{"t":"p","runs":[{"text":"If you want to engage with this beyond feeling bad about your own usage, the questions with leverage are these."}]},{"t":"p","runs":[{"text":"Is the electricity additional and clean?","b":true},{"text":" A facility powered by newly built renewable capacity is a different proposition from one that buys certificates while drawing from a coal-heavy grid. Ask about hourly matching rather than annual offsetting, because a grid at 3am is not the grid at noon."}]},{"t":"p","runs":[{"text":"Who bears the grid cost?","b":true},{"text":" Read the utility filing."}]},{"t":"p","runs":[{"text":"What is the water source and what is the local stress level?","b":true},{"text":" Consumption in a water-abundant region is a different fact from the same consumption in a drought region."}]},{"t":"p","runs":[{"text":"Is there a community benefit that is not a jobs number?","b":true},{"text":" Large data centres employ few people once built. Tax abatements are often granted against employment figures that never arrive."}]},{"t":"p","runs":[{"text":"The place your influence exists is as a citizen in a planning process, not as a user counting prompts."}]}],"check":{"question":[{"text":"An operator reports that its data centres have improved power usage effectiveness from 1.4 to 1.1 and describes this as a major environmental achievement. What is the necessary follow-up question?"}],"options":[{"text":[{"text":"Whether the measurement was independently verified, as self-reports are unreliable"}],"correct":false,"tempting":[{"text":"Verification is reasonable, but the figure could be accurate and still say nothing about total impact."}]},{"text":[{"text":"Whether the gain came from better cooling or from newer chips, which last different times"}],"correct":false,"tempting":[{"text":"An interesting engineering detail that does not bear on whether the environmental claim means anything."}]},{"text":[{"text":"Whether the sites run on renewable energy, which matters far more than efficiency"}],"correct":false,"tempting":[{"text":"An important question about carbon, but it leaves the efficiency claim unexamined and misses the demand effect."}]},{"text":[{"text":"What happened to total electricity and water use, since cheaper compute draws demand"}],"correct":true,"tempting":null}],"order":[2,3,1,0],"why":[{"text":"Efficiency lowers the cost of each unit of work, and lower cost usually raises the total. Only the totals answer the claim."}]},"prev":{"slug":"open-weights-and-accountability","title":"Open weights, and what open buys you","minutes":8},"next":{"slug":"carbon-you-can-influence","title":"The choices that actually change the number","minutes":8}}}