If humans didn't need sleep, an average person would gain about 20 extra waking years over a lifetime, but no human or animal has ever been found that needs no sleep at all. Even a brainless jellyfish shows a sleep-like state. The closest real people carry rare variants in genes such as DEC2, ADRB1 and SIK3 and do well on about six hours, not zero. To go to zero, a waking brain would have to do the jobs sleep does, starting with clearing waste, which by a simple model means working about a third harder all day.
- 175,200 hours: what 8 hours a night adds up to over 60 years, about 20 years of waking time. Calculated in this post with these example inputs: 8 hours a night (the average of non-carriers in the DEC2 family) over an assumed 60 adult years.
- 6.25 hours: the average sleep of the first family found with a short-sleep gene (DEC2, 2009).
- +60%: how much the space between brain cells grows during sleep in mice, speeding waste clearance (Xie et al., 2013).
- +134 kcal a day, about 7%: the extra energy of staying awake for 24 hours (Jung et al., 2011).
- Frigatebirds sleep just 0.69 hours a day on multi-day flights, then make up for it on land.
This is a two-part question. First, are there people who already need much less sleep, and how? Second, what would it take to need none? The first part has real families and real genes. The second is an engineering problem with biology as the constraint.
Are there people who naturally need less sleep?
Yes, a very small number. Ying-Hui Fu and Louis Ptáček's lab at UC San Francisco has been finding families since 2009 with familial natural short sleep: people who sleep about six hours, wake without alarms and show no obvious harm.
In August 2009, Fu's team reported in Science a mother in her late forties and her 17-year-old daughter who slept on average 6.25 hours a night during the study, while family members without the variant averaged about 8 hours. Both carried the same point mutation in a gene called DEC2, which helps switch off other genes. When the researchers put the human variant into mice, the mice stayed active about 1.2 hours longer each day than normal mice. The variant was so rare that it did not appear in any of 250 control people the team tested.
Scientific American: Rare genetic mutation lets some people function with less sleep (2009)The team identified the DEC2-P385R mutation in a human short-sleep family, then made transgenic mice carrying it. Activity profiles and sleep recordings showed more wake time and less sleep than control mice, and a changed response to sleep deprivation. This was the first human gene shown to set sleep length. Limit: one family, and a mouse model whose sleep differs from ours.
Read the paper (free)| Gene variant | Reported | Measured human sleep | What it seems to change |
|---|---|---|---|
| DEC2-P385R | 2009, Science | 6.25 h average | A gene switch that represses other genes |
| ADRB1 variant | 2019, Neuron | 4–6 h (three-generation family) | A wake-promoting receptor in the dorsal pons |
| SIK3-N783Y | 2025, PNAS | 6.3 h (she reported 3 h) | A kinase acting at synapses; deeper delta sleep in mice |
The SIK3 case has a quiet lesson. The woman, healthy and in her seventies, said she slept about 3 hours a day. Activity recordings showed 6.3 hours. Even true short sleepers overestimate how little they sleep.
A mutation in the ADRB1 gene was found in people who need less sleep. In cells, the mutant receptor was less stable and signalled less; mice carrying it slept less. The receptor is concentrated in the dorsal pons, where ADRB1-positive neurons are active in wake and REM sleep, and activating them causes waking. Limit: a single family and mouse work; the mechanism in humans is inferred.
Read the paper (free)Does any animal live without sleep?
None has been confirmed. Sleep, defined by behaviour (reversible quiet, slower response and catch-up after deprivation), has been found in mammals, birds, insects, worms and even jellyfish.
- Jellyfish: Cassiopea, which has a nerve net and no brain, is quiet at night, responds slowly when quiet, and is sluggish the day after being kept "awake" (Nath et al., 2017).
- Frigatebirds: on flights of up to 10 days, great frigatebirds slept only 0.69 hours a day, in seconds-long bouts, sometimes with one hemisphere and sometimes both (Rattenborg et al., 2016).
- Dolphins and some birds sleep one hemisphere at a time, which looks like "never sleeping" from outside but is not.

EEG recorders were fitted to great frigatebirds flying over the ocean for up to 10 days. In flight they slept in short bouts with one or both hemispheres, totalling 0.69 hours a day, which is 7.4% of what they sleep on land. Limit: a handful of birds of one species; the costs of the deficit were not measured.
Read the paper (free)See the data as a table
| Sleep (hours per day) | |
|---|---|
| Frigatebird on a long flight | 0.69 |
| DEC2 carriers | 6.25 |
| SIK3 carrier | 6.3 |
| Frigatebird on land (derived) | 9.3 |
| DEC2 family non-carriers | 8.06 |
Source: Rattenborg et al. 2016; He et al. 2009 via Scientific American and News-Medical; Chen et al. PNAS 2025 · Frigatebird on land derived as 0.69 / 0.074 = 9.3 h. Non-carrier figure (8.06 h) as reported by UCSF in 2019.
What would you gain if you never slept?
About a quarter to a third of your life back, and a slightly bigger food bill. Here is the working.
- Hours gained. At 8 hours a night for 60 adult years: \( 8 \times 365 \times 60 = 175{,}200 \) h, or \( 175{,}200 / 8{,}760 = 20 \) years of extra waking time. Calculated in this post with these example inputs: 8 hours a night (the average of non-carriers in the DEC2 family) over an assumed 60 adult years.
- Energy cost. Jung et al. measured 24-hour energy use rising by about 134 kcal (7%) on a day without sleep, and by 32% across the night itself. Over a year: \( 134 \times 365 \approx 48{,}900 \) kcal.
- In meals. At a reference 2,000 kcal a day, that is about 24 extra days of food a year. Calculated in this post from the 134 kcal a day measured by Jung et al. and an assumed reference intake of 2,000 kcal a day (an example input): 48,900 / 2,000 = 24.5 days. Small, but sleep is clearly not just an energy-saving trick.
Seven healthy young adults lived for three days in a whole-room calorimeter: a normal day, then 40 hours awake, then recovery sleep. Staying awake raised 24-hour energy use by about 7% (about 134 kcal) and night-time use by about 32%; recovery days were about 5% lower than baseline. Limit: seven people, one deprivation, and resting lab conditions.
Read the paper (free)Use this to check what you just read: set your current sleep and the sleep you imagine needing.
What less sleep would give you, and cost
What does sleep do that you would have to replace?
At least three jobs, each with good evidence and none fully understood: clearing waste from the brain, consolidating memory, and resetting synapses. A sleepless human would need a waking way to do all three.
Job 1: clearing waste, solved
In 2013, Lulu Xie, Maiken Nedergaard and colleagues showed in mice that natural sleep or anaesthesia expands the space between brain cells by 60%, which speeds the exchange of cerebrospinal fluid and clears β-amyloid faster. This is the glymphatic system.
Using tracer diffusion and two-photon imaging in live mice, the team found that sleep and anaesthesia were associated with a 60% increase in interstitial space, much more convective exchange between cerebrospinal and interstitial fluid, and faster β-amyloid clearance. Limit: mice, and human glymphatic flow is still harder to measure directly.
Read the paper (free)Now a simple mass balance. Waste is produced at a steady rate P. Awake, the brain clears it at rate r; asleep, k times faster, for S hours a day. Over a day, clearance must equal production.
- Set S = 8 h and k = 2, an assumed doubling during sleep anchored on the 60% space increase. Then \( (16 + 16)/24 = 1.33 \): a never-sleeping brain must clear waste 33% faster every waking hour. Calculated in this post from the mass-balance equation above with these example inputs (S = 8 h; k is an assumption anchored on the 60% wider space between cells in Xie et al.): (16 + 16)/24 = 1.33.
- Sensitivity, k = 1.6. \( (16 + 12.8)/24 = 1.20 \): 20% faster. Calculated in this post from the mass-balance equation above with these example inputs (S = 8 h; k is an assumption anchored on the 60% wider space between cells in Xie et al.): (16 + 12.8)/24 = 1.20.
- Sensitivity, k = 3. \( (16 + 24)/24 = 1.67 \): 67% faster. Calculated in this post from the mass-balance equation above with these example inputs (S = 8 h; k is an assumption anchored on the 60% wider space between cells in Xie et al.): (16 + 24)/24 = 1.67.
The model is crude: waste production is not constant, and clearance is not the only job. But it turns "you need sleep" into a design number.
Job 2: memory
Sleep spindles and slow waves coordinate the replay that moves memories from the hippocampus to the cortex; the full mechanism is in how sleep consolidates memory. A sleepless brain would need to replay offline while also taking in new input, which is exactly what attention cannot do well.

Job 3: could you sleep one half at a time?
That is the dolphin answer, and it saves nothing. If each hemisphere needs S hours offline, unihemispheric sleep still totals 2S hemisphere-hours a day.
You would never be fully unconscious, but you would spend 16 hours a day running on half a brain. The frigatebird trick is different: it cuts total sleep to 0.69 hours for a few days and then pays it back on land, where it sleeps about 9.3 hours (0.69 / 0.074). That is borrowing, not eliminating.
Calculated in this post from Rattenborg et al. (2016): 0.69 hours a day in flight is 7.4% of sleep on land, so 0.69 / 0.074 = 9.3 hours.
Every animal that has been checked sleeps. The ones that seem not to are borrowing against later.

Needing about six hours is already real, in rare carriers of DEC2, ADRB1 and SIK3 variants. Needing none would require redesigning the brain so that waking tissue clears waste about 33% faster all day (if sleep doubles clearance), consolidates memory while still taking input, and resets synapses without going offline. No animal, from jellyfish to frigatebird, has been found that does this, so zero sleep is outside known biology. The prize would be about 20 extra waking years and roughly 49,000 extra kcal a year.
If you want the opposite experiment, what actually happens when people stay awake, see what if you stayed awake for 72 hours. To sleep better rather than less, the evidence-based route is CBT for insomnia.
What's new
What if next?
- What would the Gita say to an AI? Machines never sleep. The harder question is whether they will agree to be switched off.
- What if you could never forget anything? Sleep decides what you keep. Remove the filter and the maths says your head fills up.
- What if you spent a year completely alone? With no daylight or people, even your sleep schedule wanders off.
Questions people ask
Is there anyone who doesn't need sleep?
No study has documented a healthy person who needs no sleep. The shortest healthy sleepers studied, carriers of rare DEC2, ADRB1 or SIK3 variants, sleep about four to six and a half hours.
Can you train yourself to need less sleep?
There is no good evidence that you can. The natural short sleepers found so far carry rare inherited variants, and the 2025 SIK3 case shows that even they overestimate how little they sleep.
What is the DEC2 gene?
A gene for a transcriptional repressor, a protein that switches other genes down. A 2009 study found the P385R variant in a mother and daughter who slept about 6.25 hours, and mice given the variant were active about 1.2 hours longer a day.
Why do we need sleep at all?
The best-supported reasons are clearing brain waste through the glymphatic system, consolidating memories and resetting synaptic strength. Saving energy is a minor reason: staying awake costs only about 7% more per day.
Sources
- He Y. et al.: The transcriptional repressor DEC2 regulates sleep length in mammals, Science (2009)
- Harmon K.: Rare genetic mutation lets some people function with less sleep, Scientific American (2009) (source for the 6.25 h, ages, 1.2 h and 250 controls)
- Shi G. et al.: A rare mutation of β1-adrenergic receptor affects sleep/wake behaviors, Neuron (2019)
- News-Medical: UCSF scientists identify second short-sleep gene (2019) (source for the three-generation family, the 4–6 h definition and 8.06 h for DEC2 non-carriers)
- Chen H. et al.: The SIK3-N783Y mutation is associated with the human natural short sleep trait, PNAS (2025) (source for 6.3 h measured vs 3 h reported)
- Xie L. et al.: Sleep drives metabolite clearance from the adult brain, Science (2013) (source for the 60% figure)
- Jung C.M. et al.: Energy expenditure during sleep, sleep deprivation and sleep following sleep deprivation, J Physiol (2011) (source for 134 kcal, 7% and 32%)
- Rattenborg N.C. et al.: Evidence that birds sleep in mid-flight, Nature Communications (2016) (source for 0.69 h and 7.4%)
- Nath R.D. et al.: The jellyfish Cassiopea exhibits a sleep-like state, Current Biology (2017)
- Hauglund N.L. et al.: Norepinephrine-mediated slow vasomotion drives glymphatic clearance during sleep, Cell (2025)
- Ministry of Health and Family Welfare: Tele MANAS
More on sleep and the mind on the Manthan world page.
