{
  "slug": "machine-thirst",
  "title": "Machine thirst",
  "subtitle": "A scale comparison between direct human drinking water and North American data-centre water use in 2025",
  "chart_type": "comparative-bar",
  "x_label": "Scale comparator",
  "y_label": "Liters per year",
  "units": "liters_per_year",
  "scale": "log10",
  "title_note": "",
  "thesis": "The cloud is hydraulic as well as electrical: digital systems are fed through water infrastructure, not just abstract compute.",
  "caveat": "This chart compares direct drinking water to data-centre water use as a vivid scale contrast. It does not compare full human water footprints or claim that all data-centre water is consumed in the same way everywhere.",
  "caption": "The point is scale, not moral arithmetic: the machine system is physically thirsty.",
  "annotations": [
    "At a 2-liter-per-day drinking baseline, one trillion liters equals the annual drinking water of about 1.37 billion people.",
    "The comparison stays narrow on purpose: it uses direct drinking water, not total household or lifestyle footprints.",
    "Digital infrastructure remains a system of pipes, cooling loops, and local resource negotiation."
  ],
  "method_summary": "This chart keeps the comparison deliberately simple: one person drinking-water baseline, one scaled human comparator, and one regional data-centre water-use estimate. The bar scale is logarithmic because the orders of magnitude are the point.",
  "method_notes": [
    "This chart is intentionally not a total-human-water chart. It avoids household, agricultural, or lifestyle water-footprint modeling.",
    "The bars are shown on a logarithmic scale because a linear scale would make the smaller human comparators unreadable."
  ],
  "sources": [
    {
      "label": "Reuters reporting on North American data-centre water use in 2025",
      "detail": "Provides the one-trillion-liter regional annual anchor for the machine side of the chart."
    },
    {
      "label": "Direct drinking-water assumption used for comparison",
      "detail": "Set to 2 liters per person per day, or 730 liters per year."
    }
  ],
  "alt": "Log-scale horizontal bar chart showing one person annual drinking water at 730 liters, one million people annual drinking water at 730 million liters, and North American data-centre water use at one trillion liters in 2025.",
  "rows": [
    {
      "label": "One person annual drinking water",
      "group": "humans",
      "year": 2025,
      "liters": 730,
      "unit": "liters_per_year",
      "scope": "individual",
      "source": "Assumes 2 liters of direct drinking water per day",
      "row_status": "observed",
      "method_note": "This is a direct drinking-water assumption used as a scale comparator, not a total personal water-footprint estimate."
    },
    {
      "label": "One million people annual drinking water",
      "group": "humans",
      "year": 2025,
      "liters": 730000000,
      "unit": "liters_per_year",
      "scope": "population_scale",
      "source": "Assumes 2 liters of direct drinking water per day per person",
      "row_status": "modeled",
      "method_note": "This row scales the same direct drinking-water assumption to a one-million-person population."
    },
    {
      "label": "North America data-centre water use",
      "group": "machines",
      "year": 2025,
      "liters": 1000000000000,
      "unit": "liters_per_year",
      "scope": "regional_total",
      "source": "Reuters reporting on North American data-centre water use in 2025",
      "row_status": "observed",
      "method_note": "This row is a regional annual cooling and operations water-use estimate for North American data centres."
    }
  ],
  "row_status_values": [
    "modeled",
    "observed"
  ]
}