Caffeine Half-Life: How Long Caffeine Lasts, the Math, and the Cut-Off That Protects Sleep

The coffee you drink at 2 p.m. is still a quarter-strength coffee at midnight.

Operators treat caffeine as fuel and sleep as a cost centre, and the two decisions are made in different parts of the day by what feels like different people. The afternoon version of you orders the espresso. The midnight version lies awake and blames the pitch deck.

The link between them is a number: caffeine's half-life. Once you know it, the afternoon decision stops being a mood and becomes arithmetic. This article gives you the number as the clinical and research sources state it, the worked math for a normal day, why your number may differ from your co-founder's, the single best study on caffeine timing and sleep, how much is too much, and the operator protocol that follows from all of it.

Education, not medical advice. If you have a heart condition, anxiety disorder, liver disease, take medication, or are pregnant or breastfeeding, talk to your doctor about caffeine before changing anything.

What half-life means, and what caffeine's is

Half-life is the time it takes for the amount of a substance in your blood to fall by half. It is not the time until it is gone. After one half-life you carry 50 percent of the dose. After two, 25 percent. After three, 12.5 percent. The curve never quite reaches zero; it just stops mattering at some point.

The Cleveland Clinic gives the working figure most people should start with: "Caffeine has a half-life (the time it takes to decrease by half) of five hours." It spells out the consequence: "That means that after five hours, you should have 50% of the caffeine you originally consumed in your system." And it adds the caveat that matters for the second half of this article: "But in some people, it could take up to 10 hours for caffeine to exit your bloodstream."

The research literature frames the same thing as a range rather than a point. A pharmacology review in the Annals of Pharmacotherapy states that "Caffeine levels peak 30-120 minutes after oral intake and caffeine's half-life is 3-6 hours." An older review puts the spread wider still: "The half-life of caffeine ranges from 1.5 to 9.5 hours." A small 1980 study measuring serum caffeine in three healthy adults found that "The mean serum caffeine half-life for the healthy subjects was 5.7 hours." Three subjects is not a population, but the figure sits exactly where the clinical guidance does.

So the honest answer to "how long does caffeine last" is: it halves roughly every five hours in a typical adult, faster in some people, twice as slow in others. Everything else is a consequence of that.

The math of one normal day

Take a 200 mg coffee at 8 a.m. That is a large cup or a double shot, give or take; the exact milligrams in your mug depend on the bean and the pour, which is why the figure below is labelled as an illustration rather than a measurement. Apply the Cleveland Clinic's five-hour half-life and nothing else.

Time	Hours since 8 a.m. coffee	Caffeine remaining (5-hour half-life)
8:00 a.m.	0	200 mg
1:00 p.m.	5	100 mg
6:00 p.m.	10	50 mg
11:00 p.m.	15	25 mg
4:00 a.m.	20	12.5 mg

Now add the afternoon coffee most operators actually drink. A second 200 mg at 1 p.m. stacks on the 100 mg still circulating from the morning. At 6 p.m. you carry 50 mg from the first cup and 100 mg from the second: 150 mg, three quarters of a full coffee, as you sit down to dinner. At an 11 p.m. bedtime you carry 25 mg plus 50 mg: 75 mg, which is a weak cup of tea's worth of stimulant in your blood as you try to sleep.

Now run the same day for the slow metaboliser the Cleveland Clinic describes, with a ten-hour half-life. The single 8 a.m. coffee is still 100 mg at 6 p.m. and 50 mg at 4 a.m. Add the 1 p.m. cup and bedtime arrives with roughly 150 mg on board. That person is not weak-willed about sleep. They are running a different chemistry on the same schedule.

Illustrative figures, clearly: your dose, absorption and half-life all vary. The shape of the curve does not. Caffeine stacks, and the afternoon dose is the one that lands on your night.

Why your half-life is not your co-founder's

The spread from 1.5 to 9.5 hours is not noise. It has causes, and most of them are stable facts about you rather than things you did today.

A 2018 review in Pharmacological Reviews sets out the picture. "It has been known for decades that the metabolism, clearance, and pharmacokinetics of caffeine is affected by many factors such as age, sex and hormones, liver disease, obesity, smoking, and diet." The dominant variable is genetic: "The pharmacokinetics of caffeine are highly variable among individuals due to a polymorphism at the level of the CYP1A2 isoform of cytochrome P450, which metabolizes 95% of the caffeine ingested." One liver enzyme does almost all of the work, and the gene that codes for it comes in faster and slower versions.

Three of the modifiers have direct numbers attached in the sources fetched for this article.

• Oral contraceptives. In a 16-person pharmacology study using a 200 mg caffeine dose, "Women taking oral contraceptives had prolonged caffeine elimination (mean elimination half-life, 9.7 hours versus 5.0 hours in men; P =.05)." Nearly double. A 1 p.m. coffee on a 9.7-hour half-life is still half a coffee at 11 p.m.
• Smoking. The 1988 review states it flatly: "Smoking decreases the half-life of caffeine." Smokers clear caffeine faster, which is one reason people who quit often find their usual coffee intake suddenly feels like too much.
• Liver function. The same 1980 serum study that measured 5.7 hours in healthy subjects measured two patients with alcoholic liver disease at 60 hours and 168 hours. Two patients, so treat it as a demonstration of mechanism rather than a statistic. The mechanism is the point: caffeine leaves through the liver, and a compromised liver holds it for days.

Pregnancy and medications belong in the same conversation, and the sources used here handle them by referral rather than number. The FDA notes that "Certain conditions tend to make people more sensitive to caffeine's effects, as can some medications," and advises: "If you are concerned about a condition or medication, or if you're pregnant, trying to become pregnant, or breastfeeding, we recommend talking to your health care provider." That is the right instruction and this article will not improve on it.

The operator's conclusion is simple. You do not need a genetic test. You need to find your own number by observation, which the protocol below is built to do.

The study that sets the cut-off

Most caffeine advice is folklore with a confident tone. One study did the obvious experiment properly, and its finding is the anchor for the whole protocol.

Published in the Journal of Clinical Sleep Medicine in 2013, "Caffeine effects on sleep taken 0, 3, or 6 hours before going to bed" compared "a fixed dose of caffeine (400 mg) administered at 0, 3, and 6 hours prior to habitual bedtime relative to a placebo on self-reported sleep in the home." Sleep was also "monitored objectively using a validated portable sleep monitor." The participants slept in their own beds, which is the setting that matters.

The result: "a moderate dose of caffeine at bedtime, 3 hours prior to bedtime, or 6 hours prior to bedtime each have significant effects on sleep disturbance relative to placebo (p < 0.05 for all)." Six hours out was not safe. The authors' conclusion is worth quoting in full because it is the most practical sentence in the caffeine literature:

The magnitude of reduction in total sleep time suggests that caffeine taken 6 hours before bedtime has important disruptive effects on sleep and provides empirical support for sleep hygiene recommendations to refrain from substantial caffeine use for a minimum of 6 hours prior to bedtime.

Two things to hold onto. The dose was 400 mg, a large one, so the study speaks most directly to "substantial" caffeine rather than a single small cup. And six hours was the longest interval tested, not a proven safe boundary; the study simply shows harm still present at six. Run that against the half-life table: at six hours before bed, a five-hour-half-life adult still carries more than 40 percent of the dose at lights out. The sleep data and the arithmetic agree.

For a sleep system built around this, read the sleep optimisation protocol for founders. The one rule it shares with this article: the last caffeine of the day is a sleep decision, not an energy decision.

How much is too much

The dose ceiling is better established than the timing. The FDA's consumer guidance states that for most adults it "has cited 400 milligrams a day" as an amount not generally associated with dangerous negative effects, and describes that as "about two to three 12-fluid-ounce cups of coffee." The Cleveland Clinic uses the same figure: "While 400 milligrams of caffeine a day (about four cups of coffee) is considered to be relatively safe," the rest of its sentence turns on individual sensitivity. The cup counts differ because cup sizes and strengths differ, which is itself the lesson: count milligrams, not cups.

At the far end, the FDA writes: "The FDA estimates toxic effects, like seizures, can be observed with rapid consumption of around 1,200 milligrams of caffeine." That is the territory of pure caffeine powders and stacked energy products, not coffee, and it is listed here so the scale is clear.

The more useful FDA line for an operator is this one: "there is wide variation in both how sensitive people are to the effects of caffeine and how fast they eliminate it." The 400 mg figure is a population ceiling. Your working dose may be well under it. The caffeine strategy for entrepreneurs covers how to find a dose that produces output rather than jitter.

One more mechanism to know before the protocol. Caffeine raises cortisol, and daily use builds tolerance to the morning dose but not the afternoon one. The evidence on lowering cortisol covers that study; its conclusion points the same way as the sleep data. Front-load.

The operator protocol

Three rules. They follow from the half-life, the sleep study and the dose ceiling, and nothing in them requires a product.

Rule 1: Front-load the dose

1. Decide your daily caffeine budget in milligrams, at or below 400 mg, and write it down. If you do not know what your cup contains, assume the high end for that drink type until you check.
2. Take most of the budget in the first block of the day. Caffeine peaks 30 to 120 minutes after you drink it, so a coffee at the start of the Train block is working during the Think block and fading by the afternoon.
3. Pair the first coffee with the pre-training meal rather than an empty stomach if that is your pattern; the pre-workout nutrition article covers the food side.

Rule 2: Set a hard cut-off and treat it as a bedtime decision

1. Start with the evidence-based floor: no substantial caffeine within 6 hours of bedtime. For an 11 p.m. bedtime, that is 5 p.m., and the sleep study showed harm at exactly that interval, so it is the latest acceptable boundary, not the target.
2. Apex's working recommendation, which is a judgement from the half-life arithmetic rather than a tested figure, is a cut-off 8 to 10 hours before bed: 1 p.m. to 3 p.m. for an 11 p.m. bedtime. The 1 p.m. cut-off is the table above; it is what leaves roughly 75 mg or less at lights out on a five-hour half-life.
3. If you suspect you are a slow metaboliser (you feel an afternoon coffee at bedtime, or you are on oral contraceptives), move the cut-off to midday for two weeks and watch sleep onset. Your observations are the only test that matters.
4. Put the cut-off in the calendar as a recurring event. The afternoon version of you does not get a vote.

Rule 3: Run a caffeine-free deload week every quarter

1. Pick a week with no launch, no board meeting and no travel. Deload the caffeine the way you would deload training: all the way down, on purpose, for a fixed period.
2. Expect withdrawal. The clinical description is unambiguous: patients "may present with headache, fatigue, irritability, or mood changes", and the same source notes that regular use "can produce physical and psychological dependence." Plan the week knowing days two and three will be the worst. This is information about your baseline, not a verdict on your discipline.
3. Keep training, keep the morning block, keep the evening review. The point of the week is to see what your operating system does without the stimulant layer, and to reset tolerance so the normal dose works again.
4. On return, restart at half your usual budget for a week before going back to full. Most operators find the half dose is enough. That discovery pays for the week.

What to track

Three numbers, nightly, in the same journal as the Stoic evening review: milligrams consumed, time of the last dose, and minutes to fall asleep. Two weeks of those three numbers will tell you your effective half-life more reliably than any article, including this one. The habit of measuring a stimulant you take every day is also, not incidentally, good practice for measuring everything else in the business.

Frequently asked questions

How long does caffeine stay in your system?

Caffeine has a half-life of roughly five hours in a typical adult, which means half the dose is still in your blood five hours after you drink it, a quarter after ten hours and an eighth after fifteen. Research puts the range at about 1.5 to 9.5 hours depending on genetics, hormones, smoking and liver function, and some people take up to ten hours to halve a dose.

Is a 5-hour half-life the same as caffeine wearing off in 5 hours?

No. After five hours, half of the caffeine is still there. A 200 mg coffee leaves 100 mg at five hours, 50 mg at ten hours and 25 mg at fifteen. The feeling of alertness fades before the chemical does, which is why an afternoon coffee that no longer feels like anything can still disrupt sleep that night.

What time should I stop drinking coffee to protect sleep?

A controlled home study found that 400 mg of caffeine taken six hours before bed still significantly disturbed sleep, and its authors supported a minimum six-hour cut-off for substantial caffeine. Six hours is the floor, not the target. Based on half-life arithmetic, a cut-off eight to ten hours before bedtime leaves far less caffeine on board; slow metabolisers should move it earlier still.

How much caffeine per day is safe?

The FDA cites 400 milligrams a day for most healthy adults, roughly two to three 12-ounce cups of coffee, and notes that toxic effects such as seizures can appear with rapid intake of around 1,200 milligrams. Sensitivity varies widely, some conditions and medications increase it, and anyone pregnant, breastfeeding or on medication should ask their health care provider.

Why does caffeine affect me more than other people?

A single liver enzyme, CYP1A2, metabolises about 95 percent of caffeine, and the gene for it varies between people, so half-lives range from under two hours to nearly ten. Oral contraceptives roughly doubled caffeine's half-life in one study (9.7 hours versus 5.0), smoking shortens it, and liver disease can extend it dramatically. Your dose and timing have to fit your chemistry.

Know your number, then keep the appointment

Caffeine is the one performance drug almost every operator takes daily, and almost none of them have run the arithmetic. The half-life is public. The sleep study is public. The dose ceiling is public. What separates the operator who sleeps from the one who does not is whether those facts became a calendar event.

Front-load the dose. Hold the cut-off as a sleep decision. Deload once a quarter and learn what your baseline is. The body you train in the morning is the same one metabolising the espresso at midnight, and it keeps better records than you do.

If you want the daily structure that makes rules like this hold, the free 5-Day Stoic Operator Challenge installs the morning block, the evening review and the operating rhythm in five days. The caffeine cut-off fits inside it on day one.