How much water should you actually drink?
Roughly 35 millilitres per kilogram of bodyweight, plus about 500 more per hour of exercise. Around a fifth of that arrives in food, so a 70 kg person needs about 2.5 litres total and roughly 2 litres from drinks.
Updated 2026-08-24
The eight glasses rule has no source
Eight glasses of eight ounces a day — around 1.9 litres — is the most repeated hydration advice there is, and it has no identifiable origin in research.
The likeliest source is a 1945 recommendation from the US Food and Nutrition Board suggesting about 2.5 litres of water a day for adults. The very next sentence noted that most of this quantity is contained in prepared foods. That sentence disappeared, and a total requirement including food became a drinking target on top of it.
The number is not far from wrong for an average person, which is part of why it persists. What is wrong is presenting it as a fixed requirement identical for everyone. A 50 kilogram office worker in a mild climate and a 100 kilogram labourer working outdoors in summer do not have the same requirement, and no single figure serves both.
A better starting point scales with body size: roughly 35 millilitres per kilogram of bodyweight. For a 70 kilogram person that is about 2.45 litres a day in total — close to the 1945 figure, and now attached to a reason.
A fifth of it is food
Total water requirement is not the same as how much you need to drink, because food carries a substantial amount of water and it counts.
Around twenty percent of typical intake comes from food, though the share varies with diet. Fruit and vegetables are mostly water — cucumber, lettuce, watermelon and tomatoes are all above ninety percent. Cooked rice and pasta absorb water during cooking. Soup, yoghurt and milk are largely water by mass. A diet heavy in fresh food can push the food share well above a fifth; one built on dry, processed food pushes it below.
For the 70 kilogram example, that turns a 2.45 litre total into roughly 2 litres to drink and about 0.5 litres arriving with meals. Ignoring the food contribution is what makes hydration targets feel unreachable — people are told to drink their total requirement rather than the drinking portion of it.
The practical version is to treat the drinking figure as the target and the food as a buffer you do not need to track. If you eat a lot of fruit and vegetables you are already ahead of the number.
Coffee and tea count
The belief that caffeinated drinks dehydrate you — that a coffee sets you back rather than forward — is one of the more durable pieces of nutrition folklore, and it is not supported by the evidence.
Caffeine does have a mild diuretic effect, but the fluid in the drink far exceeds the additional fluid lost. A cup of coffee is overwhelmingly water, and the net contribution to hydration is positive. Studies comparing coffee against water at moderate intakes have generally found no meaningful difference in hydration status.
The effect also diminishes with habitual use. Regular caffeine consumers develop substantial tolerance to the diuretic response, so for most people who drink coffee daily it is close to a non-issue.
Alcohol is genuinely different. It suppresses vasopressin, the hormone that tells the kidneys to conserve water, so the fluid loss can exceed the volume consumed. That is a real dehydrating effect rather than a mild one, and it is a large part of why a hangover feels the way it does.
What actually moves the number
Three things change the requirement enough to matter, and they are all about loss rather than intake.
Exercise is the largest. Sweat rates vary enormously between individuals — from well under half a litre an hour to over two — but around 500 millilitres per hour of exercise is a reasonable planning figure. For sessions beyond about an hour, particularly in heat, replacing sodium as well as water matters, because drinking large volumes of plain water while losing salt in sweat can dilute blood sodium.
Heat and humidity add to it independently of exercise. A hot climate might add half a litre a day at rest, and working outdoors in very hot conditions can add a litre or more. Humidity compounds this by reducing evaporative cooling, so you sweat more to achieve the same effect.
Altitude increases loss too, through faster breathing in drier air and increased urine output during acclimatisation. Illness with fever, vomiting or diarrhoea raises requirements sharply and is the situation where inadequate intake most often becomes clinically significant.
Thirst, urine colour, and drinking too much
For healthy adults with reliable access to fluid, thirst is a reasonably good regulator. It is not perfect — it lags during hard exercise, and the response weakens with age — but it is more useful than the popular claim that feeling thirsty means you are already dangerously dehydrated.
Urine colour is the most practical everyday check. Pale straw suggests adequate hydration; dark amber suggests drinking more. The caveats are worth knowing: B vitamins turn urine bright yellow regardless of hydration, some medications and foods change the colour, and the first urine of the morning is normally darker because it has concentrated overnight.
Drinking substantially more than you need is not usually beneficial. Excess water is simply excreted, and the main effect of a very high intake is more trips to the bathroom. There is no reserve to build up.
At the extreme it is dangerous. Consuming very large volumes rapidly — several litres in a short period, particularly while sweating heavily and replacing no sodium — can dilute blood sodium enough to cause hyponatraemia, which is a genuine medical emergency. It is rare and it happens mostly at endurance events, but it is worth knowing that the risk curve has ends on both sides rather than only one.