Everyone wants a smaller number. Six hours would free up so much time, and there is always a story about a founder or a general who runs on four. So the question feels open, like sleep need is a matter of grit or preference. It is not, really. The research here is unusually settled, and the honest answer is less flexible than most people want to hear.
For the large majority of adults, the number is seven to nine hours. That is the range where the risks of too little sleep and the diminishing returns of too much cross over. The people who truly function well on less exist, but they are a tiny fraction, and almost everyone who claims to be one is simply used to being tired.
The seven to nine range is not a guess
Major sleep bodies converge on seven or more hours per night for adults, based on a large body of work linking short sleep to worse outcomes across the board. Below that range, on average, you see more errors, slower reactions, weaker immune response, worse mood regulation, and over years, higher risk of several chronic conditions. The effects are dose-dependent, meaning the further below seven you go, the steeper the cost.
The key word is average. Your personal need might sit at seven, or at nine, and that is normal variation. What is not normal, statistically, is a real need for five. The gene variants that allow genuine short sleep are rare enough that betting you have one is a bad bet.
Duration is only half the answer
Two people can both sleep eight hours and feel completely different, because when and how consistently you sleep matters as much as how long. Your body runs on a circadian clock, and sleep taken in line with that clock is more restorative than the same hours taken at random. Shift the timing around night to night and you get a version of jet lag without leaving home, sometimes called social jet lag.
This is why a consistent bedtime and wake time often does more than adding an extra hour. A steady schedule lets the deep and REM stages fall where they should, so the same duration delivers more actual recovery. If you only change one thing about your sleep, make it the regularity, not the length.
How to find your own number
You do not need a lab. Give yourself a week or two without an alarm, on a consistent bedtime, and let your body wake naturally. The time it settles into is close to your real need. Most people are surprised it is higher than what they had been running on, because they had normalised a low-grade sleep debt.
Then watch the daytime signals, not just the clock. Needing caffeine to function, fading hard in the early afternoon, snapping at small things, or feeling foggy are all signs you are under your number. Track sleep next to your mood and energy for a couple of weeks and the pattern is usually obvious. The same tired weeks that flatten your mood tend to line up with the nights you cut short.
Frequently asked questions
Is it true some people only need four or five hours? A very small number genuinely do, due to rare gene variants. The odds you are one of them are low, and the far more common explanation for feeling fine on little sleep is being adapted to chronic tiredness rather than free of the cost.
Can I catch up on sleep at the weekend? Partly. A weekend lie-in helps with acute debt but does not fully undo a week of short nights, and swinging your schedule around adds its own social jet lag. Consistency across the week beats a big weekend recovery.
Is more than nine hours bad? Regularly needing far more than nine can be a signal worth paying attention to, sometimes pointing at poor sleep quality or an underlying issue. But the sharp risks sit on the short side, not the long side.
Does sleep quality matter more than hours? They work together. Enough hours of fragmented, badly timed sleep still leaves you short on recovery. Aim for enough duration and steady timing, and quality tends to follow.
Keep reading
- How to track your mood so it actually tells you something (mood tracking that tells you something)
- The link between sleep and overspending (how tired weeks cost you money)
- The early warning signs of burnout are hiding in your data (the signs your data sees first)