Hilbert basis theorem
Imagine you are worried that polynomials in several variables might hide infinitely complicated systems of equations — that some ideal might need infinitely many generators to pin it down. The Hilbert basis theorem is the great reassurance: it says that whenever you start from a nice, finitely-controlled ground ring and pass to polynomials, you never lose that control. Every system of polynomial conditions, no matter how it is described, can be captured by finitely many of them.
Precisely: if R is a Noetherian ring (every ideal is finitely generated, equivalently every ascending chain of ideals stabilizes), then the polynomial ring R[x] is again Noetherian. By induction R[x_1, ..., x_n] is Noetherian, and so is any finitely generated R-algebra (a quotient of such a polynomial ring). Taking R to be a field k or the integers Z, this guarantees that k[x_1, ..., x_n] and Z[x_1, ..., x_n] are Noetherian.
The proof is a clever leading-coefficient argument: given an ideal I of R[x], collect the leading coefficients of polynomials in I of each degree; these form ideals in R that are finitely generated, and lifting finitely many generators back to R[x] produces a finite generating set for I. The honest caveat is that the theorem only feeds Noetherianness forward through finite generation — infinitely many variables k[x_1, x_2, ...] is NOT Noetherian, and neither is a polynomial ring over a non-Noetherian base.
In k[x, y] the ideal (x^2, xy, y^2, x^3, ...) of all polynomials with zero constant and linear part is finitely generated — already (x^2, xy, y^2) suffices, even though one might naively list infinitely many monomials.
Finitely many generators suffice, as the theorem promises.
This theorem is the foundation that makes algebraic geometry tractable: it guarantees every affine algebraic set is cut out by finitely many polynomials, and underlies the existence of Gröbner bases used in computation.