Caffeine and Sleep: How Late Is Too Late?

The 3pm Coffee Question

You want the number. What time do I have to stop?

You've probably heard "no caffeine after 2pm," and you've probably ignored it, because you drink coffee at 4pm and fall asleep at 11 just fine. So the rule feels like something a wellness blog made up.

It isn't. But the reason it exists isn't the one you think, and "I fall asleep fine" turns out to be exactly the wrong test.

Here's the actual pharmacology, the actual trials, and a cutoff you can use.

Key Facts

• Caffeine's half-life in an average adult is about 5 hours — half of what you drank is still in you five hours later (StatPearls)
• The enzyme that clears caffeine shows up to 60-fold variation between people (Journal of Translational Medicine, 2024)
• 400 mg of caffeine taken 6 hours before bed still cut objectively measured sleep — and the participants' own sleep diaries missed it (JCSM, 2013)
• In a 2025 crossover trial, 100 mg had no significant effect even 4 hours before bed, while 400 mg cut deep sleep at 4, 8, and 12 hours out (SLEEP, 2025)
• Across 24 studies, caffeine cut total sleep time by about 45 minutes and sleep efficiency by 7% (Sleep Medicine Reviews, 2023)
• The FDA cites 400 mg a day as an amount not generally linked to negative effects in healthy adults (FDA)

What Caffeine Is Actually Doing Up There

Caffeine does not give you energy. It hides your tiredness from you.

While you're awake, a molecule called adenosine builds up in your brain. Think of it as a tiredness meter that fills all day and drains while you sleep. The more adenosine that's accumulated, the sleepier you feel. Researchers call it "an endogenous sleep-regulatory substance," which is a formal way of saying your brain makes its own sleeping pill [^1].

Caffeine's molecular shape is close enough to adenosine that it fits into the same receptors — the docking ports adenosine uses to deliver its message. Caffeine plugs into all four adenosine receptor subtypes without triggering them, and blocking the A2a type in particular is what produces the wakefulness [^2].

So the tiredness signal is still there. Caffeine is just standing in the doorway so the message can't get delivered. About 30 minutes after you drink it, caffeine reaches your brain and starts blocking [^1].

Which explains a familiar experience. When it wears off, the tiredness doesn't arrive — it was already there, waiting, and all of it lands at once.

Half-Life: Why the 3pm Cup Is Still Working at Midnight

Here's the part that changes people's minds.

Caffeine's half-life in an average adult is about 5 hours [^2]. Half-life means the time it takes your body to clear half of what's in you. Not all of it. Half.

Run the math on a 200 mg coffee at 3pm:

At 1am — deep into your night — you still have the caffeine equivalent of a decent cup of tea circulating. And your body absorbs nearly all of what you drink; oral bioavailability is close to 100% [^2].

That's the whole mechanism. Caffeine isn't uniquely evil. It just hangs around far longer than the "alert" feeling does, and the alert feeling is the only part you notice.

What the Controlled Trials Actually Show

Three studies, in order of importance.

Drake 2013: the 0, 3, and 6 hour study

Twelve healthy adults took 400 mg of caffeine — a big dose, roughly a large strong coffee — at bedtime, 3 hours before bedtime, or 6 hours before bedtime, versus placebo. Double-blind, randomized, each person doing every condition [^3].

Sleep was measured two ways: a self-report diary, and a validated at-home sleep monitor (not a full lab sleep study, which matters).

The objective monitor found significant reductions in total sleep time at every timing, including 6 hours before bed [^3]. Caffeine at 5pm measurably shortened sleep that started at 11pm.

Small study, 12 people, one dose size. Take the direction seriously and the exact minute counts with a grain of salt.

Gardiner 2025: dose matters as much as timing

Twenty-three healthy young men, placebo-controlled, double-blind, randomized crossover, seven conditions: placebo plus 100 mg and 400 mg of caffeine at 12, 8, and 4 hours before bed, measured with in-home sleep-lab equipment [^4].

The results split cleanly by dose:

• 100 mg (roughly a small coffee): no significant effect on objective or subjective sleep, at any timing — including 4 hours before bed [^4].
• 400 mg at 4 hours out: total sleep time down 50.6 minutes, deep N3 sleep down 29.7 minutes [^4].
• 400 mg at 8 hours out: N3 down 15.3 minutes [^4].
• 400 mg at 12 hours out: N3 down 20.6 minutes [^4].

Read that again. 400 mg at 11am measurably reduced deep sleep that night. The authors' own recommendation: 100 mg is fine up to 4 hours before bed, but avoid 400 mg as a single dose within 12 hours of bedtime [^4].

Also 23 young men, so don't treat these as universal constants.

The meta-analysis: the average picture

Pooling 24 studies: caffeine reduced total sleep time by about 45 minutes, cut sleep efficiency by 7%, added about 9 minutes to how long it took to fall asleep, and added about 12 minutes of wake time after falling asleep. Light N1 sleep went up; deep N3 and N4 sleep went down [^5].

Their practical recommendation, converted to a normal cup: a coffee at 107 mg per 250 mL (about 8.5 fl oz) should be finished at least 8.8 hours before bed, and a standard pre-workout serving at 217.5 mg at least 13.2 hours before bed [^5].

An earlier systematic review reached the same conclusion from a wider base: caffeine lengthens sleep latency, cuts total sleep time and sleep efficiency, and worsens how good sleep feels [^6].

The Thing Almost Everyone Gets Wrong

You are not a reliable witness to your own sleep.

In the Drake study, both the diary and the objective monitor caught the disruption when caffeine was taken at bedtime or 3 hours before. But at 6 hours before bed, only the objective measure caught it ¹. People reported sleeping fine. They didn't.

The authors flagged exactly why this matters: the subjective-versus-objective gap is worst when the awakenings are short and scattered, and they named it as a likely reason people ignore caffeine advice in the first place ¹.

The 2025 trial shows the same shape from another angle. At 8 and 12 hours before bed, 400 mg of caffeine cut deep sleep without making it harder to fall asleep ². You'd feel nothing. Your night would still be worse.

So "I sleep fine after evening coffee" isn't evidence you sleep fine. It's evidence you can't feel the difference — a different claim. If you want to know how long you're actually taking to drop off, we covered that in sleep latency.

Why Your Friend Drinks Espresso at 9pm and Sleeps Like a Rock

They might not be lying. Caffeine metabolism varies enormously between people.

Caffeine is broken down in the liver, mostly by an enzyme called CYP1A2 ³. How much of that enzyme you have and how fast it works is partly genetic — and the range is wild. A 2024 systematic review reports up to 60-fold variation between individuals in the rate of caffeine breakdown ⁴. Twin studies back the genetic part: identical twins matched each other far more closely than fraternal twins did ⁴.

Some people are fast metabolizers. Others are slow metabolizers, hold onto caffeine longer, and get more of everything — more alertness, more jitters, more sleep disruption. The same review notes that people whose sleep suffers after caffeine are more likely to be slow metabolizers ⁴.

A second gene matters too. Variation in ADORA2A, which codes for one of the adenosine receptors caffeine blocks, is linked to caffeine-driven anxiety ^{5 4}. So sensitivity is both about how fast you clear it and how strongly your receptors respond.

It isn't only genetics. Smoking cuts caffeine's half-life by up to half, and pregnancy — especially the last trimester — can stretch it out to as long as 15 hours ³. Reviews also note that older adults' sleep may be more sensitive to caffeine than younger adults' ⁶.

Practical upshot: population averages are a starting point, not your answer. You are allowed to be an outlier in either direction. You just have to test it honestly instead of by vibes.

How Much Caffeine Is Actually in Your Drink

Here's the FDA's table. Note the unusual thing about it: every row is normalized to 12 fluid ounces, so the numbers are directly comparable to each other rather than to whatever cup you happen to own ⁷.

Source: U.S. Food and Drug Administration ⁷

Two things jump out.

The coffee number is higher than the one in your head. A big 12 oz cup of strong brew can land near 250 mg on its own. The FDA's own framing: 400 mg a day, their stated limit for healthy adults, is "about two to three 12-fluid-ounce cups of coffee" ⁷.

The ranges are enormous. Energy drinks span 41 to 246 mg for the same volume. Two cans that look identical in the cooler can differ by a factor of six. Read the label.

For reference, the meta-analysis modeled a standard coffee at 107 mg per 250 mL and a pre-workout serving at 217.5 mg ⁸ — roughly two coffees in one scoop.

So What's the Actual Cutoff?

Combining all of it, here's what I'd tell a friend.

If you drink a lot (300–400 mg in one hit): stop 10 to 12 hours before bed. That's the pre-workout scoop, the double espresso, the big cold brew. The 2025 trial found deep-sleep effects still measurable 12 hours out at 400 mg ².

If you drink a normal cup (roughly 100 mg): a 4-hour buffer looks fine. That's the one clean piece of good news in the 2025 data — 100 mg showed no significant effect on sleep even at 4 hours before bed ².

If you're picking one number: about 8 to 9 hours before bed for a normal coffee. That's the meta-analysis recommendation ⁸, and it's the safest single rule. For an 11pm bedtime, that's roughly 2pm. The old advice was right, it just never explained itself.

Then test yourself, honestly. Cut evening caffeine for two weeks and compare against two weeks with it. Not one night against one night — sleep is noisy enough that a single comparison tells you nothing.

And if you're using caffeine to paper over chronic short sleep, the caffeine isn't the problem to fix.

Where Komori Fits

Komori is being built as a contactless device that senses from above the bed — no wearable, no camera. When it ships, it's designed to report sleep position, movement and restlessness, bed-exit events, and your room's conditions overnight: temperature, humidity, CO2, light, and noise. It's a general wellness device, not a medical or diagnostic one, and it doesn't measure anything about caffeine.

Where it's meant to help with a question like this is the boring part: a multi-night record you didn't have to remember to keep. The Drake study's central finding was that people's own diaries missed the disruption. A two-week comparison beats a one-night impression, and what your overnight movement says about your sleep is one of the few things a device across the room can observe directly rather than infer.

Not a caffeine detector. Just a record that doesn't flatter you.

This post is educational, not medical advice. If caffeine is tangled up with a heart condition, an anxiety disorder, a pregnancy, or a medication you take, that's a conversation with your doctor, not with a blog.

---

Footnotes

1. Drake C, Roehrs T, Shambroom J, Roth T. "Caffeine Effects on Sleep Taken 0, 3, or 6 Hours before Going to Bed." Journal of Clinical Sleep Medicine, 2013. ↩ ↩²

2. Gardiner C, et al. "Dose and timing effects of caffeine on subsequent sleep: a randomized clinical crossover trial." SLEEP, 2025. ↩ ↩² ↩³

3. Evans J, Richards JR, Battisti AS. "Caffeine." StatPearls Publishing, updated 2024. ↩ ↩²

4. Low JJL, et al. "Genetic susceptibility to caffeine intake and metabolism: a systematic review." Journal of Translational Medicine, 2024. ↩ ↩² ↩³ ↩⁴

5. Reichert CF, Deboer T, Landolt HP. "Adenosine, caffeine, and sleep-wake regulation: state of the science and perspectives." Journal of Sleep Research, 2022. ↩

6. Clark I, Landolt HP. "Coffee, caffeine, and sleep: A systematic review of epidemiological studies and randomized controlled trials." Sleep Medicine Reviews, 2017. ↩

7. U.S. Food and Drug Administration. "Spilling the Beans: How Much Caffeine is Too Much?" ↩ ↩² ↩³

8. Gardiner C, et al. "The effect of caffeine on subsequent sleep: A systematic review and meta-analysis." Sleep Medicine Reviews, 2023. ↩ ↩²