← Reference · Nestor G Pestelos Jr
Mathematics · Applied Mathematics
Exponential Growth
Reference entry · last updated August 24, 2026
Exponential growth is a pattern of change in which a quantity's rate of increase is proportional to its current size, so the quantity grows by a constant percentage of itself each period rather than by a constant amount. A quantity N growing exponentially at rate r follows \( N(t) = N_0 (1+r)^{t} \) in discrete time, or \( N(t) = N_0 e^{kt} \) in continuous time. The defining feature is that growth compounds on top of itself: each period's increase is added to a larger base than the last, which is what makes the curve accelerate rather than run in a straight line.
Exponential growth vs. linear growth
Linear growth adds a fixed amount each period: \( N(t) = N_0 + ct \). Exponential growth adds a fixed percentage of the current amount each period. Early on, the two can look similar, because a percentage of a small number is itself small. The difference becomes stark once the base has grown: a fixed 10% of 100 is 10, but a fixed 10% of 10,000 is 1,000. The absolute increase itself keeps growing, which a linear process never does. Plotted on a standard (linear) y-axis, exponential growth produces a curve that bends sharply upward; plotted on a logarithmic y-axis, it becomes a straight line, which is why log scales are the standard tool for reading growth rates off a chart.
Doubling time and the Rule of 72
A useful way to characterize exponential growth is by its doubling time: how long it takes the quantity to double, which stays constant regardless of the starting value. Solving \( N_0(1+r)^t = 2N_0 \) for t gives the exact doubling time \( t = \ln(2) / \ln(1+r) \). For everyday mental arithmetic, this is commonly approximated by the Rule of 72: divide 72 by the growth rate (as a percentage) to estimate the number of periods to double.[1]
| Growth rate | Rule-of-72 estimate | Exact doubling time |
|---|---|---|
| 2% / period | 36.0 periods | 35.0 periods |
| 6% / period | 12.0 periods | 11.9 periods |
| 9% / period | 8.0 periods | 8.0 periods |
| 12% / period | 6.0 periods | 6.1 periods |
The constant 72 is a rounded stand-in for \( 100 \times \ln(2) \approx 69.3 \), chosen because 72 divides evenly by 1, 2, 3, 4, 6, 8, 9, and 12, which makes the mental division easier at a small accuracy cost. The approximation is tightest in the 6–10% range; two adjacent variants exist for other ranges: the Rule of 70, more accurate near 2% and commonly used for population growth rates, and the Rule of 69.3, exact for continuous compounding.[1] The identical arithmetic runs in reverse for shrinking quantities: at a 6% annual rate, purchasing power halves in about 72 / 6 = 12 years.
Why intuition fails
Because the absolute size of each period's increase depends on the current base, and that base itself keeps growing, most of an exponential curve's total increase happens in its last few doublings. That makes the growth look deceptively slow for most of its length and then suddenly urgent. Physicist Albert Bartlett illustrated this with a bacteria-in-a-bottle thought experiment: bacteria that double every minute, placed in an empty bottle at 11:00 a.m., fill the bottle completely by noon. One minute before the end, at 11:59, the bottle is only half full. Five minutes before the end, at 11:55, only 3% of the bottle is full. 97% still looks like open space.[2] Bartlett used the example to argue that a population or a resource consumption rate growing at a constant percentage does not look alarming until it is almost exhausted, since most of the growth curve's length is spent looking nearly empty.
Real-world examples
Compound interest and inflation both follow the exponential form directly: a balance or a price level growing at a fixed annual percentage rate compounds on itself period over period. In epidemiology, the early stage of an outbreak (before immunity, distancing, or intervention change the contact rate) approximates exponential growth, with the reproduction number determining the rate. In computing, Gordon Moore's 1965 observation that the number of components achievable at minimum manufacturing cost per integrated circuit was increasing "at a rate of roughly a factor of two per year" is the origin of what became known as Moore's Law: an extrapolated exponential trend, read off three years of cost-curve data, that predicted 65,000 components per circuit by 1975.[3] A single bacterium dividing every 20 minutes under ideal lab conditions is the standard biology-textbook case of unconstrained exponential population growth.
Why it doesn't continue forever
Exponential growth requires that the growth rate stay constant regardless of the quantity's size. That assumption requires unlimited resources, and it essentially never holds indefinitely in a physical system. Population ecology's standard correction is the logistic growth model, which adds a term for the environment's carrying capacity, K: \( \frac{dN}{dt} = r_{max} N \frac{(K-N)}{K} \). When the population N is small relative to K, the added factor is close to 1 and growth is nearly indistinguishable from exponential; as N approaches K, the factor shrinks toward zero and growth is throttled down by resource competition and waste accumulation, producing an S-shaped curve rather than the unbroken J-shape of pure exponential growth.[4] Exponential growth is best understood as the special case of logistic growth where the carrying capacity is effectively infinite: a useful local approximation for the early stage of growth, valid only until a real system starts to feel its limits.
See also
- ELI5: What Is Exponential Growth? — a picture-book explainer covering the same ground
- Reference: Exponential Decay — the mirror-image process this entry grows away from
- Reference: Exponential Backoff — a specific engineering technique that deliberately uses exponential growth (of a retry delay) to relieve system contention
References
- ^ "Rule of 72," Wikipedia — https://en.wikipedia.org/wiki/Rule_of_72
- ^ Albert A. Bartlett, "Arithmetic, Population and Energy" (lecture transcript) — https://www.albartlett.org/presentations/arithmetic_population_energy_transcript_english.html
- ^ Gordon E. Moore, "Cramming More Components onto Integrated Circuits," Electronics, vol. 38, no. 8, April 19, 1965 (reprinted Proceedings of the IEEE, vol. 86, no. 1, January 1998) — https://www.cs.utexas.edu/~fussell/courses/cs352h/papers/moore.pdf
- ^ "45.3 Environmental Limits to Population Growth," Biology 2e, OpenStax — https://openstax.org/books/biology-2e/pages/45-3-environmental-limits-to-population-growth