What Is Pi and Why Does It Never End?
Pi is not 3.14 rounded badly. Its digits never repeat because no fraction can equal it — and that was proved, not assumed.
The short answer
Pi is the ratio of a circle circumference to its diameter, identical for every circle. Its decimal expansion never ends and never repeats because pi is irrational: it cannot be written as a ratio of two integers. This was proved by Johann Lambert in 1761. Pi is also transcendental, proved by Ferdinand von Lindemann in 1882, meaning it is not the root of any polynomial with integer coefficients, which is why squaring the circle with compass and straightedge is impossible.
Transcript
What is pi, and why does it never end? Start with what it actually is, because the definition does the work.
Take any circle. Measure the distance around it, the circumference. Divide by the distance across it, the diameter. You get pi. Always. A coin, a stadium, the orbit of a planet — the ratio is identical. That constancy is what makes pi a constant of geometry rather than a measurement of any particular circle.
Now, why do the digits never end? Because pi is irrational. That is a precise claim: it cannot be written as a fraction of two whole numbers. And here is the key fact — any fraction, when you write it as a decimal, either terminates or eventually repeats forever. One third is 0.333 repeating. Seven elevenths repeats too. If pi ever settled into a repeating pattern, it would be a fraction. It is not, so it cannot.
This was proved in 1761 by Lambert. It is not that we ran out of computing power. No fraction exists.
And pi goes further. It is transcendental — not the solution to any polynomial equation with whole-number coefficients. That is why you cannot square the circle.
Test yourself on Geometry
5 questions, easy to hard. No account needed to try it.
How is pi defined?
What does it mean for a number to be irrational?
Why does being irrational guarantee that the digits of pi never repeat?
Who first proved that pi is irrational, and roughly when?
Pi is transcendental. What classical problem does that fact prove impossible?
The longer answer
Most people meet pi as a number to memorise, which is the least interesting thing about it. Pi is not fundamentally a decimal at all. It is a ratio, and the reason its decimal expansion behaves so strangely is a consequence of what kind of number that ratio turns out to be.
Begin with the definition, because it carries more weight than it appears to. Take any circle, measure its circumference, measure its diameter, and divide. The result is pi. The striking part is the word any. A circle the size of a coin and a circle the size of a planetary orbit give the identical ratio. This is not an experimental finding that happens to hold so far; it follows from the fact that all circles are geometrically similar — every circle is a scaled copy of every other. Scaling multiplies circumference and diameter by the same factor, so their ratio is untouched. Pi is a fact about the shape, not about any particular circle.
From that definition the familiar formulas are immediate rather than arbitrary. Circumference equals pi times diameter, which is the definition rearranged, and equals 2 pi r since the diameter is twice the radius. The area formula, pi r squared, needs a short argument — slice the disc into thin wedges and rearrange them into an approximate rectangle of height r and width half the circumference — but it is the same constant appearing again.
Now the question of the digits. The precise claim is that pi is irrational: there are no whole numbers p and q with pi equal to p over q. The endless non-repeating decimal is not the definition of irrationality; it is a consequence of it, and understanding the link is what the harder exam questions test.
Here is the link. Write out any fraction as a decimal by long division. At each step you have a remainder, and for a denominator q there are only q possible remainders — 0 through q minus 1. Keep dividing and you must eventually hit a remainder you have seen before, at which point the whole sequence of digits from that point repeats. So every fraction produces a decimal that either terminates, when the remainder hits zero, or falls into a repeating cycle. One third gives 0.333 repeating; one seventh gives a six-digit block that repeats forever; three quarters terminates at 0.75.
That argument runs in both directions. A decimal that terminates or eventually repeats can always be converted back into a fraction. So the terminating-or-repeating decimals are exactly the rational numbers. If the digits of pi ever settled into a repeating block, pi would be a fraction — and it is proved that it is not. The digits therefore cannot repeat, and since they do not terminate either, they run on forever without pattern.
That proof matters, and its date matters. Johann Lambert proved pi irrational in 1761. This is not a case of computation not having gone far enough, or of measurement being imprecise. It is a theorem: no fraction equals pi, and no amount of future computing will find one. Archimedes, around 250 BCE, had already bounded pi between 223/71 and 22/7 using inscribed and circumscribed 96-sided polygons — a brilliant piece of work, but it established approximation, not irrationality. The common classroom habit of calling 22/7 the value of pi confuses the two: 22/7 is a convenient approximation that diverges from pi in the third decimal place.
Pi has a second, stronger property. It is transcendental, proved by Ferdinand von Lindemann in 1882. Transcendental means it is not a root of any polynomial equation with integer coefficients. The contrast makes this clearer: the square root of two is irrational, but it satisfies x squared minus 2 equals 0, so it is algebraic. Pi satisfies no such equation of any degree.
That result closed a problem that had stood for over two thousand years. Squaring the circle asks for a square with the same area as a given circle, constructed using only compass and straightedge in finitely many steps. It can be shown that every length constructible by those tools is algebraic — it must satisfy some polynomial with integer coefficients. A square of area pi r squared needs a side of length r times the square root of pi. Since pi is transcendental, so is its square root, so the length is not constructible. Lindemann did not fail to find a construction; he proved that none exists. The same line of reasoning also settles doubling the cube and trisecting a general angle as impossible.
Two loose ends are worth tidying up, because both are commonly misunderstood.
First, an endless non-repeating expansion does not mean the digits are random or that every possible string appears somewhere. Whether pi is a normal number — one in which every finite digit sequence occurs with the expected frequency — is unproven. The digits look statistically random in every test run so far, but looks are not proof, and claims that your phone number must appear somewhere in pi are conjecture rather than fact.
Second, pi is not merely a geometric curiosity. It appears throughout mathematics in places with no visible circle: in the normal distribution of statistics, in Fourier analysis, in Euler's identity linking it to e and i, and in the infinite series that sum to pi squared over six. The reason is that pi is really the constant of periodicity and rotation, and anything that oscillates or repeats eventually brings it into view. That is also how pi is computed today — not by measuring circles, which no amount of physical precision could take beyond a few decimals, but by evaluating rapidly converging infinite series.