Running a human brain costs roughly 20% of the body’s total resting energy, despite the organ weighing only about 2% of body mass.1PubMed Central. Appraising the brain’s energy budget That imbalance, a ten-fold premium per gram compared to an average tissue, makes the brain one of the most expensive organs in biology. But that headline figure barely scratches the surface. Where exactly those calories go, how the bill changes across a lifetime, what evolutionary compromises our species made to afford it, and what happens when the energy supply is cut off are all part of the real accounting.
Where the Energy Actually Goes
Most of your brain’s fuel is burned on communication between neurons, not on thinking per se. Roughly half of the brain’s total energy expenditure goes to powering excitatory synapses, the junctions where neurons pass signals to each other using the neurotransmitter glutamate.2Current Biology. Neuronal energy use and brain evolution An early detailed model of the gray matter’s energy budget broke the costs down further: action potentials (the electrical impulses that travel along nerve fibers) consumed about 47% of signaling energy, while the postsynaptic effects of glutamate took about 34% and the maintenance of cells’ resting electrical state used around 13%.3PubMed. An energy budget for signaling in the grey matter of the brain Recycling glutamate after it has been released accounted for only about 3%.
A later analysis put the numbers slightly differently, estimating that pre- and postsynaptic mechanisms together consume about 55% of the energy spent on signaling, or about 41% of the cortex’s total energy if you include the housekeeping tasks that keep cells alive, like building proteins and trafficking materials around.4Neuron. How Much Does a Brain Cost? The Real Numbers Either way, the takeaway is the same: the synapse is the single most expensive component of brain operation. Sending signals is cheap compared to receiving, processing, and cleaning up after them.
This matters because the brain contains somewhere around 86 billion neurons and a comparable number of non-neuronal cells.5PubMed Central. The remarkable, yet not extraordinary, human brain as a scaled-up primate brain and its associated cost Each neuron forms thousands of synaptic connections. The total number of synapses runs into the hundreds of trillions, and each one is a tiny metabolic furnace burning fuel every time it fires. That staggering connectivity is what makes you conscious, creative, and capable of reading this sentence, and it is also what makes the brain so phenomenally expensive to operate.
Fueling the Machinery
Under normal conditions, the brain runs almost exclusively on glucose. But neurons themselves are not great at managing their own fuel supply; they get substantial help from astrocytes, star-shaped support cells that outnumber neurons in some brain regions. A framework called the astrocyte-neuron lactate shuttle describes how astrocytes take up glucose, partially metabolize it, and then pass the resulting lactate to neighboring neurons as a ready-to-burn fuel.6PubMed Central. Brain energy homeostasis: the evolution of the astrocyte-neuron lactate shuttle hypothesis The two cell types express different versions of the enzyme that handles lactate, which keeps the fuel flowing in the right direction, from astrocyte to neuron.
This shuttle is not just a background logistics detail. Experiments have shown that disrupting lactate transport between astrocytes and neurons causes amnesia, specifically blocking long-term memory formation.7Cell. Regulation of Learning and Memory by Astrocyte-Neuron Lactate Transport When researchers knocked out the transporters that move lactate out of astrocytes, injecting lactate directly into the affected brain region rescued the animals’ ability to form lasting memories. Knocking out the transporter on the neuron side, however, could not be rescued by lactate alone, because the neuron had lost the ability to import it. The implication is striking: memory is not just an electrical or molecular coding event. It is also, at its root, a metabolic event that depends on a working fuel pipeline.
The Childhood Peak
If you think the adult brain’s 20% share of the body’s energy is extreme, consider what happens in childhood. The brain’s energy demands do not peak at birth, when relative brain size is largest, but several years later, during early childhood. In children around ages four to five, the brain consumes about 66% of the body’s resting energy expenditure.8PubMed Central. Metabolic costs and evolutionary implications of human brain development Two-thirds of a child’s resting calories are going to a single organ.
This peak coincides with the period of most intense synaptic growth and refinement. Young brains are building connections at a furious rate, and each new synapse costs energy not just to form but to maintain and test. The timing has another consequence that has attracted research interest: the years of highest brain energy demand are also the years just before the “adiposity rebound,” the age when childhood body-fat levels bottom out before climbing again toward puberty. One hypothesis proposes that the developing brain’s enormous energy appetite actively suppresses fat deposition during this period, and that variation in that appetite may influence later obesity risk.9PubMed Central. A hypothesis linking the energy demand of the brain to obesity risk
The bill drops during adolescence, partly thanks to synaptic pruning. The teenage brain eliminates excess synaptic connections it built during childhood, keeping the strong ones and discarding the weak. Because every synapse costs fuel to maintain, trimming them reduces the brain’s overall energy demand and contributes to the increases in cognitive efficiency seen during the teenage years.10PubMed Central. Adolescent Neurodevelopment By adulthood, the brain has settled back to its 20% baseline, running a leaner, more optimized network.
What Evolution Traded Away
No species gets a brain this expensive for free. Evolution had to pay for human encephalization, and the currency was other tissues and life-history strategies. The “expensive tissue hypothesis” proposes that organisms with larger brains compensate by shrinking other metabolically costly organs. Studies of Lake Tanganyika cichlid fish, which show brain-size variation comparable to primates, have found a clear negative association between brain size and gut size after controlling for ecology and shared ancestry.11PubMed Central. Comparative support for the expensive tissue hypothesis: Big brains are correlated with smaller gut and greater parental investment in Lake Tanganyika cichlids A bigger brain came paired with a smaller digestive tract and greater investment in fewer offspring.
Humans followed a version of the same playbook. Compared with other primates of similar body size, humans are notably “under-muscled,” carrying less skeletal muscle and more body fat, particularly during infancy.12Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physiology. Metabolic correlates of hominid brain evolution That extra fat serves as a stored energy reserve to feed the brain when caloric intake dips, while the reduced muscle mass lowers the body’s competing energy demands. The arrangement is especially visible in human infants, who are born with proportionally far more body fat than any other primate newborn, precisely because their brains are already burning through an outsized share of incoming calories.
There are additional costs on the maternal side. Growing a large-brained fetus is so energetically demanding that the rate of fetal brain growth appears constrained by the mother’s metabolic capacity.13PubMed Central. Maternal investment, life histories, and the costs of brain growth in mammals Research suggests that human gestation length is limited not by the size of the birth canal, as was once assumed, but by the mother’s ability to sustain the metabolic output required to grow the fetus.14PubMed Central. Metabolic hypothesis for human altriciality In other words, human babies are born “early” and helpless not because their heads would not fit through the pelvis otherwise, but because their mothers have hit a metabolic wall.
The Neuron-Count Bottleneck
The energetic cost of the brain scales directly with the number of neurons it contains. This is not a rough correlation; the relationship is linear.15PubMed Central. Metabolic constraint imposes tradeoff between body size and number of brain neurons in human evolution Every neuron you add comes with a predictable price tag in calories. For great apes eating raw diets, the hours available for foraging and the low caloric density of unprocessed plants impose a hard ceiling on how many neurons can be supported. A gorilla, despite being much larger than a human, has far fewer brain neurons because it cannot afford to fuel a bigger brain on its diet and feeding schedule.
Whether cooking unlocked the human brain’s expansion is a question that has generated heated debate. One prominent analysis argued that learning to cook starchy foods freed up enough calories to break the ape bottleneck. However, a counter-analysis using a similar mathematical framework found that the predictable number of brain neurons in the hominin lineage varied much more with foraging efficiency than with body mass, and that archaeological evidence of fire control does not correlate with the timeline of brain expansion.16PubMed Central. Human Brain Expansion during Evolution Is Independent of Fire Control and Cooking Experiments in mice in the same study showed that cooking meat did not increase its caloric availability. The question is far from settled, but the simple “cooking made us smart” story may need revising.
Not every species solves the energy problem the same way. Mormyrid electric fishes in Africa have evolved extreme encephalization rivaling that of primates, but rather than trading away gut or muscle tissue, they appear to simply run a higher overall metabolic rate. Brain size in these species correlates directly with resting oxygen consumption, suggesting they pay for their large brains by burning more total energy rather than redistributing a fixed budget.17PubMed Central. The costs of a big brain: extreme encephalization results in higher energetic demand and reduced hypoxia tolerance in weakly electric African fishes The tradeoff for them shows up differently: they are more vulnerable to low-oxygen environments.
When Glucose Runs Low
The brain is essentially glucose-dependent under normal conditions, but it has a backup plan. During prolonged fasting, the liver converts fatty acids into ketone bodies, which cross the blood-brain barrier and can supply a substantial fraction of the brain’s fuel.18PubMed Central. Effects of Ketone Bodies on Brain Metabolism and Function in Neurodegenerative Diseases People born with genetic defects that prevent ketone production still develop normally, which suggests ketones mainly serve as an emergency fuel for periods of illness or extended food deprivation rather than being critical for day-to-day brain function.19PubMed. Cerebral ketone body metabolism
The brain also has an aggressive strategy for protecting its glucose supply when the body is under stress. Through a mechanism called “cerebral insulin suppression,” the stressed brain activates the sympathetic nervous system to suppress insulin secretion from the pancreas.20Frontiers in Neuroenergetics. How the selfish brain organizes its supply and demand Because muscles and fat cells need insulin to absorb glucose, this effectively diverts blood sugar away from the body and toward the brain, which imports glucose through a transporter that does not require insulin. Researchers have described this as the “selfish brain” principle: under threat, the brain prioritizes itself.21PubMed Central. The selfish brain: stress and eating behavior When the stress system misfires chronically, the resulting metabolic disruption may contribute to overeating and weight gain, because the body’s peripheral tissues are persistently starved of glucose even when calories are abundant.
Does Thinking Harder Cost More?
One of the most counterintuitive findings in brain energetics is how little extra fuel hard mental work requires. During episodes of intense cognitive effort, local changes in blood flow and glucose use rarely exceed 5% above the brain’s baseline resting consumption.1PubMed Central. Appraising the brain’s energy budget The brain at rest is not idling. It is running a massive baseline operation: maintaining resting potentials across billions of neurons, sustaining spontaneous synaptic activity, and supporting the default-mode network that is active during daydreaming, self-reflection, and mind-wandering. Focused thought reshuffles where that energy goes rather than dramatically increasing the total.
This is part of why the popular notion that you can “burn extra calories” by doing Sudoku or studying for an exam is mostly wrong. Your brain’s metabolic rate is dominated by the enormous fixed costs of keeping the network alive and ready. The incremental cost of solving a math problem on top of that baseline is a rounding error in your daily caloric budget. You might feel mentally exhausted after a long exam, but the fatigue is more about neurotransmitter depletion and attentional resource management than about having burned significantly more fuel.
The Heat Problem
Every joule of energy the brain uses ends up as heat. Unlike muscles, which convert metabolic energy into mechanical work, the brain performs no physical labor. All of its energy expenditure ultimately becomes thermal waste.22PubMed Central. Brain temperature and its role in physiology and pathophysiology: Lessons from 20 years of thermorecording On average, each gram of brain tissue releases about 0.66 joules of heat per minute.23Frontiers in Neuroscience. Brain temperature and its fundamental properties: a review for clinical neuroscientists Scale that up to the whole organ and the brain is essentially a biological space heater sealed inside an insulating skull.
The cooling system is blood flow. Cerebral circulation carries heat away from active tissue and distributes it. Interestingly, when a brain region becomes more active during a task, the local increase in blood flow typically outpaces the increase in oxygen consumption, which means the activated region actually cools slightly even as it works harder. Meanwhile, regions that become less active during the task may warm up because their blood flow drops. The net effect is that blood flow shields the brain from dangerous temperature swings by preventing hot spots from spreading.24PubMed Central. Theoretical model of temperature regulation in the brain during changes in functional activity This coupling of blood flow to activity is so reliable that it is the foundation of functional brain imaging. Both PET and fMRI measure neural activity indirectly, by detecting changes in local blood flow or oxygenation that accompany it.25PubMed Central. Behind the scenes of functional brain imaging: a historical and physiological perspective
What Happens When the Power Goes Out
The brain’s extreme energy appetite comes with a devastating vulnerability: it has almost no energy reserves. Unlike muscle, which stores glycogen for emergency use, the brain holds only enough glucose and oxygen for a few minutes of normal function. When blood flow is interrupted, as in a stroke, energy production fails rapidly, and the consequences cascade within seconds.26PubMed Central. Reperfusion brain injury: focus on cellular bioenergetics Neurons lose the ability to maintain their resting electrical potentials, uncontrolled ion flow triggers toxic levels of neurotransmitter release, and the mitochondria that generate most of the cell’s energy become damaged. Even restoring blood flow creates a secondary wave of injury as oxygen floods back into compromised tissue.
Conditions like epilepsy illustrate a different kind of energy crisis. Seizures involve massive, synchronized neuronal firing that dramatically spikes local energy demand. The interplay between metabolic disruption and neuroinflammation can become self-reinforcing, with each exacerbating the other and creating a cycle that makes seizures more likely and harder to control.27PubMed Central. The interplay between metabolism and neuroinflammation in epilepsy: mechanisms and therapeutic perspectives Understanding brain energy use has become central to epilepsy research, with interventions like ketogenic diets, which shift the brain toward ketone-based metabolism, showing clinical benefit precisely because they alter the brain’s fuel supply.
The Brain Versus the Data Center
An increasingly common comparison pits the brain’s energy efficiency against that of modern artificial intelligence. The human brain runs on roughly 20 watts, comparable to a dim light bulb. The latest AI systems trained to perform tasks that humans handle effortlessly, like recognizing faces or holding a conversation, consume orders of magnitude more power during both training and operation. A recent analysis of what it would take to build an artificial superintelligence concluded that the fundamental barrier is energy: any system more intelligent than a human brain but built with current or near-future hardware would require so much power that it would be impractical to run.28PubMed Central. The energy challenges of artificial superintelligence
The brain achieves its efficiency because it computes in a fundamentally different way than silicon chips. Neurons are both the processor and the memory, computation happens at the synapse where signals are transmitted, and the whole system runs at millivolt scales with wet chemistry rather than etched circuits. The fact that biology had three and a half billion years to optimize energy use while digital computing has had a few decades puts the comparison in perspective, but it also frames an aspirational engineering target. Neuromorphic chips, designed to mimic the brain’s architecture, are being pursued partly because the brain has proven that intelligence at 20 watts is physically possible.
What Happens While You Sleep
The brain does not shut down during sleep, and its energy consumption drops only modestly. But sleep serves a critical maintenance function related to the costs of running such an energy-intensive organ. During wakefulness, metabolic activity produces waste products that accumulate in the spaces between brain cells. The glymphatic system, a network that uses cerebrospinal fluid to flush these wastes, operates far more efficiently during sleep. In mouse experiments, glymphatic clearance increased by about 80-90% during sleep compared to wakefulness, and the amount of protein waste cleared roughly doubled.29PubMed Central. The Sleeping Brain: Harnessing the Power of the Glymphatic System through Lifestyle Choices Slow-wave sleep in particular drives this process, with slow oscillatory brain activity pumping cerebrospinal fluid through interstitial spaces more effectively.
This suggests that sleep is, at least in part, the price of running a 20-watt organ nonstop. The brain generates metabolic debris proportional to its energy use, and without adequate sleep to clear it, the waste accumulates. The proteins removed during sleep include amyloid-beta, a molecule implicated in Alzheimer’s disease, which has led researchers to investigate whether chronic sleep deprivation increases dementia risk by impairing this cleanup process. The expensive brain, it seems, requires expensive maintenance.