If gcd(a, n) = 1, then a^φ(n) ≡ 1 (mod n), where φ is Euler's totient function. This generalizes Fermat's Little Theorem (where n = p gives φ(p) = p−1) and is essential for understanding RSA cryptography and computing modular exponentiation.
You already know from the totient function that φ(n) counts how many integers from 1 to n share no common factor with n — the units modulo n. Euler's theorem makes a powerful claim about what happens when you raise any such unit to the φ(n) power: you always land on 1. Think of this as a "reset" — repeated multiplication eventually cycles back to the identity.
To build intuition, consider n = 10 and a = 3. We have φ(10) = 4 (the units are 1, 3, 7, 9). Compute: 3¹ = 3, 3² = 9, 3³ = 27 ≡ 7 (mod 10), 3⁴ = 81 ≡ 1 (mod 10). The power cycle resets after exactly 4 steps. Euler's theorem guarantees this will happen for any a coprime to n — it might reset earlier (when the order divides φ(n)), but it always resets by step φ(n).
You may recognize a special case: if n = p is prime, then φ(p) = p − 1, and the theorem says a^(p−1) ≡ 1 (mod p) for any a not divisible by p. This is exactly Fermat's Little Theorem. Euler's theorem is the generalization that works for composite moduli too, which is why it has broader applications. The proof strategy is elegant: the set {a·1, a·3, a·7, a·9} (multiplying all units by a) is just a rearrangement of {1, 3, 7, 9} modulo n. Multiplying all elements on both sides yields a^φ(n) times the product of units ≡ that same product — cancel, and you get a^φ(n) ≡ 1.
The main application you'll encounter is modular exponentiation for large numbers. If you need to compute a^k mod n, you can reduce the exponent: a^k ≡ a^(k mod φ(n)) (mod n) when gcd(a, n) = 1. This is the mathematical engine behind RSA encryption, where message decryption requires computing m^d mod n for astronomically large numbers. Euler's theorem is what makes that computation tractable — without it, you could never decrypt quickly.