A cube root releases one copy of a prime for every THREE under the sign, where a square root needs two. 54 is 2 × 3³, so the three 3s come out as a single 3 and the 2 stays: ∛54 = 3∛2. And unlike a square root, a negative number has a real answer — ∛−27 is −3, because multiplying three negatives leaves a negative.
Why a negative has a real cube root
(−3) × (−3) is 9, and 9 × (−3) is −27. Three negatives leave a negative, because they cannot pair off — one is always left over.
That is the whole difference an odd index makes. An even root of a negative has no real value because any real number squared is positive; an odd root has one for exactly the opposite reason.
So ∛−27 = −3 with no complex numbers involved, and ∛−54 = −3∛2. A tool that refuses these has reused square-root logic where it does not apply.
Three roots, not one
Every non-zero number has three cube roots, evenly spaced 120° apart on a circle in the complex plane. One is real; the other two are a conjugate pair.
They are not mysterious. The other two are always the real root multiplied by (−1 + i√3)/2 and (−1 − i√3)/2 — the primitive cube roots of unity, which are the same two numbers whatever you started with.
“The” cube root means the real one by convention. That convention matters in Cardano’s formula, where taking the principal cube root at each step does not always give the real answer you wanted — the pairing between the two cube roots has to be respected.
Threes, not pairs
The simplification rule is the same shape as for square roots with one number changed: a cube root needs THREE of a prime to release one, not two.
So 8 = 2³ comes out whole, while 4 = 2² does not come out at all — two 2s are not enough. That is the difference that makes ∛4 irrational where √4 is 2.
It also means perfect cubes are much rarer than perfect squares. There are 31 squares below 1000 and only 10 cubes, so most cube roots you meet will be irrational.
Doubling the cube
One of the three classical Greek construction problems: given a cube, build one with twice the volume using only straightedge and compass. That needs a length of ∛2 times the original.
It is impossible, and the reason is exactly the kind of fact theRational Zeros Calculator establishes: x³ − 2 has no rational root, so it is irreducible over the rationals, and a degree-3 irreducible polynomial has no root constructible with straightedge and compass — those build only degrees that are powers of two.
The proof took two thousand years to find and is about a page long. ∛2 ≈ 1.2599 is perfectly computable; it is just not constructible.
Cube roots in real use
Volume to side. A tank holding 1000 litres as a cube is 10 units on a side. Any question that starts from a volume and wants a length is a cube root.
Scaling. Doubling every length multiplies volume by eight; to double the volume you multiply lengths by ∛2 ≈ 1.26. That gap is why a scale model’s weight is so far from what its size suggests.
Solving cubics. Cardano’s formula is built from two cube roots, which is why the Cubic Equation Calculator exists and why it has to be careful about which cube root it takes.
How the decimal is found
The same Newton iteration as for square roots, with the weights changed: the new guess is (2x + n/x²) / 3 rather than (x + n/x) / 2. The 2 and the 3 come from the index.
It still converges quadratically, so the correct digits roughly double each step and seven steps cover the sixty places this page will print.
As on the square-root page, the iteration runs on whole numbers scaled by a power of ten rather than on floating point, so nothing is lost past the seventeenth digit — and the result is checked by cubing it back.
Sources and methodology
The results here are classical; these are where they are set out.
Method. The radicand is factorised and every complete triple of a prime is pulled out. The decimal is computed separately by Newton iteration on integers scaled by a power of ten. Sign is handled before either: an odd index preserves it, so the whole calculation runs on the absolute value and the sign is restored at the end. That engine is verified on every change against 68 hand-written assertions, including that every simplified radical at indices two to five multiplies back to its radicand exactly, and that ∛−27 returns −3 rather than refusing. The count and the per-case breakdown are published on the formula verification page.
Read the guide
Square roots, where a negative has no real answer, are on the Square Root Calculator. For a coefficient in front or a higher index, the Simplify Radicals Calculator handles the general case.