How ATP Works: The Energy Currency of Every Living Cell

At any moment your body holds only a small pool of ATP, commonly estimated at tens of grams. Yet an average adult regenerates on the order of 50 kilograms of it every day, roughly 100 moles, which is about body weight. That mismatch is the key to the whole subject. ATP is not stored energy in the way fat or a battery is. It is a short-lived carrier that cells spend and rebuild many times a minute. This article explains what ATP is, how a molecule of glucose is converted into about 30 to 32 of them, and why the final step works like a turbine.

Educational content, not medical advice. Talk to your doctor before changing any medication, treatment, or routine. Information here reflects current research but every person's situation differs.

What is ATP, and what is it made of?

ATP stands for adenosine triphosphate. The name is a parts list:

• Adenine, a nitrogen-containing base (the same "A" found in DNA and RNA).
• Ribose, a five-carbon sugar. Adenine plus ribose together is called adenosine.
• Three phosphate groups linked in a chain, which is the "triphosphate."

When a cell needs energy, an enzyme splits off the outermost phosphate. That reaction, called hydrolysis (splitting with water), turns ATP into ADP (adenosine diphosphate) plus a free phosphate ion (written Pi) and releases usable energy: ATP + H₂O → ADP + Pi + energy. The reverse reaction, attaching a phosphate back onto ADP, is how the cell recharges. It requires an energy input, and that input comes from breaking down food.

A short history of the molecule

ATP was discovered in 1929, when Karl Lohmann in Germany (working with Jendrassik), and independently Cyrus Fiske and Yellapragada Subba Rao in the United States, isolated it from muscle. In 1941 the biochemist Fritz Lipmann proposed that ATP serves as the go-between linking energy-releasing reactions to energy-requiring ones, an idea that established its reputation as the cell's "energy currency." Source: Adenosine triphosphate, Wikipedia, which summarizes the history.

The currency analogy is useful but worth stretching. Cells do not hoard ATP the way people save money. They earn and spend it in continuous small transactions, more like a poker chip passed back and forth at a busy table than a gold bar in a vault.

Why does breaking ATP release energy?

Many textbooks say energy is "stored in the high-energy bond" between the phosphates. Chemists and biochemists generally treat that wording as a shortcut, and it causes a common misunderstanding. Breaking any chemical bond costs energy. The net release happens because the products of the reaction, ADP and phosphate, are more stable than ATP and water. Several factors contribute. The three phosphate groups each carry negative charge and repel one another, so the crowded chain is strained. Once the end phosphate is free, the products are better stabilized by water and by resonance (electrons spreading across more than one arrangement).

A useful analogy is a compressed spring held shut by a catch. The energy is not "in the catch." It lies in the strained arrangement of the whole spring, and releasing the catch lets it relax.

The numbers

Under defined laboratory conditions (called standard conditions), hydrolysis of ATP to ADP and phosphate releases about 30.5 kilojoules per mole, or 7.3 kilocalories per mole. Inside a living cell, where concentrations of ATP, ADP, and phosphate are far from standard, the figure is larger, around 57 kilojoules per mole (roughly 13.6 kilocalories per mole). Source: Wikipedia.

The cell keeps ATP far from equilibrium with ADP and phosphate, rather like holding water at the top of a hill. That is why the reaction releases so much energy when allowed to proceed.

How the energy actually gets used: coupling

A cell cannot simply let ATP release heat, because heat does not turn a motor or build a molecule. Instead, enzymes couple the ATP reaction to a reaction that needs energy. A typical case is the first step of glycolysis, where an enzyme transfers a phosphate from ATP onto glucose. The phosphate-bearing sugar is now more reactive and trapped in the cell. The energy-releasing reaction and the energy-requiring one are physically linked on the same enzyme, like two weights on opposite ends of a pulley. The heavy side (ATP hydrolysis) lifts the light side (the otherwise unfavorable reaction).

Because ATP can donate a phosphate to thousands of different proteins and small molecules, one currency powers wildly different jobs.

Job	What ATP does	Example
Movement	Powers motor proteins	Myosin pulling on actin filaments in muscle
Transport	Drives pumps against a concentration gradient	The sodium-potassium pump, an ATP-using enzyme that maintains ion balance in the cell
Construction	Activates building blocks	Synthesis of proteins, DNA, and lipids
Signaling	Adds phosphates to proteins to switch them on or off	Kinase enzymes (enzymes that attach phosphates)

The sodium-potassium pump (Na⁺,K⁺-ATPase) was discovered by Jens Skou, who shared the 1997 Nobel Prize in Chemistry for the work, according to the Nobel Prize press release summary. For the muscle example in more detail, see our guide to how muscles work.

Stage 1: glycolysis splits sugar in the cytoplasm

The cell's main ATP-making program starts with glucose and has three stages. The first is glycolysis, a ten-step sequence of reactions that happens in the cytoplasm, not in the mitochondria. It is found in nearly every living organism, which suggests it is ancient and predates the arrival of mitochondria in eukaryotic cells.

Glycolysis splits one glucose (six carbons) into two pyruvate molecules (three carbons each). The accounting is a start-up loan: the cell invests 2 ATP in the early steps to prime the sugar, then earns 4 ATP in the later steps, for a net gain of 2 ATP. It also converts two molecules of NAD⁺ into NADH, a carrier of high-energy electrons that matters a great deal later.

Glycolysis does not need oxygen

Glycolysis runs without oxygen. If oxygen is scarce, cells can keep glycolysis going by regenerating NAD⁺ through fermentation. In muscle during intense effort, pyruvate is converted to lactate. In yeast, it becomes ethanol and carbon dioxide. The yield is only 2 ATP per glucose, but the pathway is fast and works when oxygen cannot be delivered quickly enough. Red blood cells, which have no mitochondria, depend on it.

The link reaction: preparing for the Krebs cycle

When oxygen is available, pyruvate enters the mitochondria. There, a large enzyme complex converts each pyruvate to acetyl-CoA, a two-carbon unit attached to coenzyme A. In this step:

• One carbon leaves as carbon dioxide. This is one source of the CO₂ you exhale.
• One NADH is produced.
• The remaining two-carbon fragment is attached to coenzyme A, forming acetyl-CoA.

Per glucose, two pyruvates go through this step, so the totals are 2 CO₂ and 2 NADH.

Stage 2: the Krebs cycle strips off electrons

The Krebs cycle (also called the citric acid cycle or tricarboxylic acid cycle) runs in the mitochondrial matrix, the fluid-filled interior. It is named for Hans Krebs, who with William Johnson identified it in 1937 at the University of Sheffield and received the 1953 Nobel Prize in Physiology or Medicine. Other researchers, including Albert Szent-Györgyi, worked out parts of it in the 1930s. Source: Citric acid cycle, Wikipedia. For a deeper walk-through of each reaction, see our dedicated explainer on the Krebs cycle.

Think of the cycle as a wheel that picks up a two-carbon package at the top and drops off its carbon atoms as CO₂ on the way around. Acetyl-CoA (two carbons) combines with oxaloacetate (four carbons) to form citrate (six carbons). A series of enzyme-catalyzed reactions then strips away carbon as CO₂ and harvests electrons, regenerating oxaloacetate to start again. That regeneration is why it is a cycle rather than a line.

Each turn of the cycle yields:

• 3 NADH
• 1 FADH₂ (another electron carrier)
• 1 GTP or ATP (a directly made energy molecule, with the type varying by tissue)
• 2 CO₂ released

One glucose produces two acetyl-CoA units, so the cycle turns twice per glucose: 6 NADH, 2 FADH₂, and 2 GTP or ATP.

The direct ATP profit is small. The point of the Krebs cycle is not the ATP it makes itself but the electron carriers it fills. They are like loaded cargo trucks, and the real payoff comes at the unloading dock.

Stage 3: oxidative phosphorylation makes most of the ATP

Oxidative phosphorylation takes place on the inner mitochondrial membrane and has two linked parts: an electron transport chain and an ATP-making enzyme.

The electron transport chain pumps protons

NADH and FADH₂ hand their electrons to a series of protein complexes embedded in the inner membrane, called Complexes I through IV. As electrons move from one complex to the next, they lose energy in small steps, like water falling down a staircase rather than off a cliff. The complexes use that energy to pump protons (hydrogen ions, H⁺) from the matrix into the space between the inner and outer membranes.

The pumping is not equal at all steps. According to the standard accounting summarized in Wikipedia's P/O ratio article, Complexes I, III, and IV move about 4, 4, and 2 protons per pair of electrons, which adds up to about 10 protons for each NADH. FADH₂ enters at Complex II, which does not pump protons, so it feeds fewer pumps and yields about 6 protons. That is why NADH is worth more ATP than FADH₂.

The result is a proton gradient: a higher concentration of H⁺ outside the matrix than inside, and an electrical charge difference across the membrane. Picture the pressure behind a dam.

Oxygen is the final electron acceptor

At the end of the chain, the spent electrons combine with oxygen and protons to form water:

4e⁻ + O₂ + 4H⁺ → 2H₂O

This is why we need oxygen. It is not burned directly. It is the sink that keeps electrons flowing. Without it, the chain backs up, NADH cannot be unloaded, and the upstream pathways stall. Toxins such as cyanide are dangerous in part because they block this final step in the chain.

ATP synthase: the molecular turbine

Protons flow back down the gradient into the matrix through a channel in ATP synthase, an enzyme that works like a turbine in a hydroelectric dam. The flow spins part of the enzyme, a ring of proteins called the c-ring, and that rotation drives a central shaft inside the enzyme's head, which squeezes ADP and phosphate together to make ATP.

Some specifics are known with confidence. Each full turn of the shaft produces three ATP. Animal mitochondria have a c-ring with eight subunits (reported by John Walker's group in 2010, as described in a 2023 Plant Physiology commentary), so eight protons pass through per turn. Dividing 8 protons by 3 ATP gives about 2.7 protons per ATP. Bringing ADP and phosphate into the mitochondrion and ATP out costs roughly one more proton, so about 3.7 protons are needed per ATP that reaches the cytoplasm. Source: Wikipedia's P/O ratio article. Other organisms have different ring sizes, with bacteria and chloroplasts reported to use 10 to 15 subunits, which means a higher proton cost per ATP.

The scientists behind the idea

The proton-gradient mechanism was a radical proposal. In 1961, Peter Mitchell suggested that ATP is made by an electrochemical gradient across a membrane rather than by a direct chemical intermediate. He received the Nobel Prize in Chemistry in 1978 for this chemiosmotic hypothesis, according to the Nobel committee. Paul Boyer proposed how the enzyme makes ATP, and John Walker determined its structure, confirming Boyer's mechanism; they shared half of the 1997 Nobel Prize in Chemistry. Source: Nobel Prize press release, 1997. The idea that an enzyme physically rotates to make ATP was, when first proposed, one of the strangest claims in biochemistry, and it has since been confirmed by direct observation of rotation in the enzyme's head.

How many ATP does one glucose really yield?

Here the textbooks have changed their numbers over the decades, and the reason is itself instructive.

Older books gave 36 or 38 ATP per glucose. Those numbers assumed that each NADH always produces exactly 3 ATP and each FADH₂ exactly 2. Measurements and the structure of ATP synthase suggest the true values are closer to 2.5 and 1.5, because protons are also used for transport and because some leak across the membrane. A university course page notes that the old 38-ATP maximum is "merely an estimate," and that much of the energy in glucose is lost as heat. Source: Penn State open education page.

With the modern conversion factors, the accounting looks like this:

Stage	Direct ATP or GTP	Electron carriers	ATP from carriers (approx.)	Subtotal (approx.)
Glycolysis	2	2 NADH	3 to 5 (depends on the shuttle that moves electrons into the mitochondria)	5 to 7
Link reaction	0	2 NADH	5	5
Krebs cycle (2 turns)	2	6 NADH, 2 FADH₂	15 + 3 = 18	20
Total	4	10 NADH, 2 FADH₂	26 to 28	about 30 to 32

Two points explain the range. First, the NADH made in the cytoplasm during glycolysis cannot cross the inner mitochondrial membrane, so cells use "shuttle" systems to move its electrons in. Different shuttles deliver the electrons at different points, so they are worth different amounts of ATP. This varies by tissue. Second, the 2.5 and 1.5 values are conventions and theoretical calculations, not constants. The P/O ratio, meaning the ATP made per oxygen atom used, is still discussed in the literature, with some sources favoring 2.5 and others a little higher.

Reading the ledger. Roughly 4 ATP are made directly, in glycolysis and the Krebs cycle. About 26 to 28 more come from the electron carriers through oxidative phosphorylation, which is why most of the harvest depends on oxygen.

Compare that range with the 2 ATP that glycolysis alone provides under fermentation. Oxygen multiplies the harvest from the same glucose by a factor of 15 or so. That is the biochemical reason animals depend on breathing, and why aerobic life could support larger, more energetic bodies. For a look at how the mitochondrion itself came to exist, see how mitochondria became cell powerhouses.

What about fat, protein, and ketones?

Glucose is the textbook fuel, but cells burn several others. All feed into the same machinery, mostly through acetyl-CoA.

Fuel	How it enters	Notes
Fatty acids	Chopped two carbons at a time by beta-oxidation into acetyl-CoA	A 16-carbon fatty acid, palmitate, nets roughly 106 ATP with modern conversion factors (older textbooks say 129; exact values depend on the conversion factors used)
Amino acids	Converted to pyruvate, acetyl-CoA, or Krebs cycle intermediates	Used more heavily when other fuels are scarce
Ketone bodies	Made by the liver from fatty acids, then converted back to acetyl-CoA in other tissues	A fuel the brain can use during prolonged fasting or very low carbohydrate intake

The difference between 106 and 129 for palmitate, like the difference between 30 and 38 for glucose, comes from the conversion factors used for NADH and FADH₂ rather than from different chemistry. Fat stores far more energy per gram than carbohydrate, roughly 9 kilocalories per gram compared with about 4. Fat carbons are more reduced, meaning they carry more electrons to hand over, and fat is stored without the water that glycogen (stored glucose) carries along. To see how your body manages the glucose side of this, read how the body stores and releases glucose, and for what happens during fasting, see our review of what the science says about intermittent fasting.

How do cells keep ATP supply matched to demand?

Cells do not make ATP at a fixed rate. They make it as needed, and the control system is built into the chemistry.

The key signal is the ratio of ATP to ADP. When a muscle contracts, it converts ATP to ADP quickly, and the rising ADP speeds up the electron transport chain and ATP synthase. When ATP is plentiful, ADP is scarce, the proton gradient builds up, and the chain slows. This feedback, called respiratory control, is like a thermostat that opens the furnace when the room cools. Several enzymes in glycolysis and the Krebs cycle also respond to the ATP and ADP levels, so the supply pathways speed up and slow down in step.

Muscle adds a second buffer, a compound called phosphocreatine that can quickly pass a phosphate to ADP to rebuild ATP during short bursts of effort. And because the pool of ATP is small, a cell whose supply is cut off, for instance by a poison that blocks the electron transport chain, runs into trouble very quickly.

A worked example: what does 100 moles of ATP per day cost?

The turnover figure is easier to believe once you do the arithmetic. Take the round estimate of about 100 moles of ATP regenerated per day, and use two numbers from earlier in this article: about 30 ATP per glucose, and about 57 kilojoules of free energy per mole of ATP under cellular conditions.

• Glucose needed. 100 mol of ATP divided by 30 ATP per glucose is about 3.3 mol of glucose. With a molar mass of 180 grams per mole, that is roughly 600 grams of glucose per day if glucose were the only fuel. In reality the fuel mix includes fats and other sources, and the figure is meant only to show scale.
• Energy handled. 100 mol multiplied by 57 kJ per mol is about 5,700 kilojoules, or roughly 1,360 kilocalories of free energy passing through the ATP system each day. That is the same order of magnitude as the food energy a person eats, which makes sense, since ATP is the intermediary between the two.

Notice the mismatch between the 600 grams of glucose and the 50 kilograms of ATP. The ATP is not made from scratch each time. If the body holds somewhere around 50 to 100 grams at once, a common range of estimates, then each molecule must be rebuilt several hundred times a day, on the order of 500 to 1,000, to reach 50 kilograms. That repeated rebuilding is the practical meaning of "currency." It also connects to a basic principle of thermodynamics (see thermodynamics and entropy): no conversion is perfectly efficient, so a good deal of the energy in food leaves the body as heat, which is part of how we stay warm. And enzymes, the protein catalysts that make each step fast enough to matter, are what let cells run these reactions at body temperature; our explainer on how catalysts speed up reactions covers the principle.

Common confusions about ATP

"ATP is where the body stores energy"

Not in the long-term sense. The body stores energy as glycogen and fat. ATP is the quickly turned-over spending currency, and because its pool is small compared with its turnover, each molecule is recycled many times a day.

"Mitochondria are the only place ATP is made"

Mitochondria make most of the ATP in cells that have them, but glycolysis in the cytoplasm makes some, and cells without mitochondria, such as mature red blood cells, rely on it. Plants add another source: chloroplasts use light energy to build ATP during photosynthesis, a process covered in how photosynthesis works.

"Exactly 30 ATP per glucose"

The figure is an estimate with a range of about 30 to 32 in current textbooks. Actual yields vary by tissue, by the shuttle used, and with how leaky the membrane is.

"Energy is stored in the phosphate bond"

As described above, the bond-breaking itself costs energy. The release comes from the more stable products. The shorthand is common, but it hides how the reaction really works.

"More ATP is always better"

Cells match production to demand. A high ATP level slows its own production, and in some tissues, such as brown fat, cells deliberately let protons leak across the membrane to make heat instead of ATP. In that case, the gradient's energy is being deliberately spent as warmth.

The takeaway

ATP is best understood as a transaction, not a store. A cell holds only a small pool, spends it by transferring a phosphate to whatever needs work done, and pays it back by burning food, with the final payment made by a spinning enzyme driven by a proton gradient. Glycolysis, the Krebs cycle, and the electron transport chain are three installments of one process that turns the chemical energy in glucose into a form every cell can spend.

The most useful detail may be the least obvious: the biggest payoff does not come from any reaction that makes ATP directly. It comes from loading electrons onto carriers, running them down a staircase of protein complexes, and letting protons flow back through a turbine. That is why oxygen matters so much, why the exact ATP count is a range rather than a single number, and why interrupting any link in the chain, from the fuel supply to the final electron handoff, shows up so quickly in a living body.

Key Takeaways

• ATP is the universal energy currency of cells. Hydrolysis to ADP and phosphate releases about 30.5 kJ/mol under standard conditions and around 57 kJ/mol under typical cellular conditions.
• Glycolysis splits glucose into pyruvate in the cytoplasm for a net gain of 2 ATP and 2 NADH, without needing oxygen.
• The link reaction and the Krebs cycle strip carbon off as CO₂ and load electrons onto NADH and FADH₂ in the mitochondrial matrix.
• The electron transport chain pumps protons across the inner mitochondrial membrane. ATP synthase uses the proton flow to make ATP.
• With oxygen, one glucose yields about 30 to 32 ATP in current accounting, compared with 2 from fermentation. About 4 are made directly and the rest through the electron carriers.

FAQ

Q: Why is ATP called the energy currency of the cell? A: Like money, ATP is a standard unit that many different processes accept. Cells store energy in glycogen and fat but spend it as ATP, using enzymes that transfer a phosphate group from ATP to power movement, transport, construction, and signaling.

Q: How much ATP does the body actually make? A: Commonly cited estimates put daily ATP turnover for an average adult at around 50 kilograms (roughly 100 moles), comparable to body weight, even though the body holds only a small amount at once. Each molecule is recycled many times a day.

Q: What happens to ATP production when oxygen runs out? A: Without oxygen, the electron transport chain stops because oxygen is the final electron acceptor. Cells fall back on glycolysis and fermentation, regenerating NAD⁺ by converting pyruvate to lactate or ethanol. That keeps some ATP flowing but yields only 2 per glucose rather than about 30 to 32.

Q: Why do we breathe out carbon dioxide? A: Carbon dioxide is the carbon waste of burning fuel. It is released in the link reaction, where pyruvate becomes acetyl-CoA, and in the Krebs cycle, where the remaining carbon atoms are stripped off. The carbon in the CO₂ you exhale was once part of the food you ate.

Q: What does NADH do? A: NADH is an electron carrier. It collects high-energy electrons from glycolysis, the link reaction, and the Krebs cycle and delivers them to the electron transport chain, where their energy is used to pump protons. Each NADH is worth roughly 2.5 ATP in current accounting.

Q: Why is the ATP yield per glucose given as a range? A: The exact figure depends on how many protons ATP synthase needs per ATP, how many are lost to leakage and transport, and which shuttle carries cytoplasmic NADH electrons into the mitochondria. Current textbooks give about 30 to 32. Older ones gave 36 to 38 using rounder conversion factors.

Q: What makes mitochondria the powerhouse of the cell? A: Mitochondria have a highly folded inner membrane, with the electron transport chain and ATP synthase embedded in it. The folding increases the surface area available, and the double membrane creates a separate compartment in which a proton gradient can build up.

Sources

• Wikipedia contributors, "Adenosine triphosphate" (discovery history; free energy of hydrolysis; daily turnover estimate). Accessed October 19, 2026. Figures for standard and cellular free energy are widely reproduced in biochemistry textbooks such as Berg, Tymoczko, and Stryer, Biochemistry.
• Wikipedia contributors, "Citric acid cycle" (history; per-turn yields; location).
• Wikipedia contributors, "P/O ratio" (protons pumped per complex; protons per ATP; P/O values).
• Nobel Prize Outreach, "The Nobel Prize in Chemistry 1997, press release" (Boyer, Walker, Skou); Mitchell's 1978 Nobel Prize in Chemistry for the chemiosmotic hypothesis.
• Yokoyama R. "EngineeRING the gear of ATP synthase: a way to address complexity and flexibility of photosynthesis." Plant Physiology. 2023;192(2):691-693. PMC10231449. Cites Watt IN, et al. "Bioenergetic cost of making an adenosine triphosphate molecule in animal mitochondria." PNAS. 2010;107(39):16823-16827.
• Penn State Department of Biology, "Oxidative phosphorylation and ATP yield", open educational resource.

---

Read next:

• How Cells Divide: Mitosis, Meiosis, and the Cell Cycle Explained
• How Photosynthesis Works: Light, Dark Reactions, and the Calvin Cycle
• How the Lymphatic System Works: Immunity, Lymph Nodes, and Cancer