Hold two numbers side by side. One is the small pour of dark liquid in front of you, a few gulps, gone in minutes. The other is a figure more than a thousand times larger, hidden entirely from view, most of it spent thousands of miles away before the beans were ever roasted.
The gap between those two numbers is the whole story here, and it is stranger than it first looks.
The liquid you drink is tiny. The water it took to get there, on one widely quoted estimate, is not. We want to walk through where that big number comes from, why almost none of it happens in your kitchen.
First, the small number. Chapagain and Hoekstra noted that “A standard cup of coffee is 125 ml.” Other definitions run higher, up to a US 8-ounce mug of around 240 ml but whether the cup holds 125 or 240 millilitres, it is a rounding error next to the water hiding behind it.
Now the big number. It comes from a 2003 report and a 2007 peer-reviewed follow-up by Ashok Chapagain and Arjen Hoekstra. Their original report put it plainly: they wrote that “We found that for drinking one standard cup of coffee in the Netherlands we need about 140 litres of water, by far the largest part for growing the coffee plant.”
The peer-reviewed version landed on the same figure. In Ecological Economics, the authors reported that a standard cup of coffee in the Netherlands costs about 140 litres of water, while tea costs about 34 litres. Because their tea serving was twice the volume of their coffee serving, coffee required roughly eight times as much water per unit of beverage volume.
To make the scale vivid, the authors offered a ratio: they calculated that “we need more than 1100 drops of water for producing one drop of coffee.” More than a thousand drops in, one drop out.
This is one modelled estimate, based on Dutch coffee consumption and trade in the late 1990s. The authors nevertheless said the Dutch figures were quite representative of the global average, so they could be cited more generally.
Why almost all of it is spent on the farm, not in the kitchen
When you picture the “water in coffee,” you probably picture the kettle. But the brewing water, and even the water used to wash and pulp the beans at the mill, barely registers. Almost the entire footprint is attributed to growing the plant.
How lopsided is it? Reviewing the numbers, Paul Hicks noted that more than 99% of coffee’s water footprint is the water used for growing the coffee plant. The original study calculated that water used in wet processing amounted to only about 0.34% of the water used to grow it. By the time water reaches your cup, the vast bulk of coffee’s calculated water footprint has already been associated with a field abroad, much of it as rain falling on coffee plants.
That last detail is the key to reading the number honestly, and it’s where the figure gets contested.
What the figure does and doesn’t tell you
The 140 litres lumps together kinds of water that behave very differently. Water-footprint methodology distinguishes “green water,” the rainwater stored in soil and used by plants; “blue water,” surface water or groundwater; and “gray water,” a measure associated with water pollution. Adding them into a single headline number is exactly what some critics object to.
Hicks made this critique directly. As he put it, the nuance gets lost in the simple take-home message that it takes 140 litres of water to make a cup of coffee. His point is that rainwater moving through a plant and returning to the atmosphere through evapotranspiration is not equivalent to groundwater pumped from an aquifer, so counting them together can flatten a distinction that matters on the ground.
And the ground is where he thinks the attention belongs. The small processing share can still be locally serious: Hicks points out that for people living downstream of coffee farms or mills, that 0.34% could represent their primary drinking-water supply. His broader argument is that “It matters a great deal where products are sourced.” Two cups with similar headline footprints can carry very different real-world water pressures depending on local rainfall, groundwater availability and pollution.
So how should we hold the number? Not as a verdict on whether to drink coffee. A litre count with no map attached can’t tell us that. What it does well is redirect attention. The interesting water in a cup of coffee was never the 125 millilitres we swallow, or even the kettle. It’s the season of weather that fell on a plant somewhere in Brazil or Colombia or El Salvador, and whether that somewhere had water to spare.
