This is easier to understand if you think about the geometric meaning of a determinant.
A matrix [itex]A[/itex] represents a linear transformation in a given basis. The determinant is a property of the matrix, but it's also a property of the linear transformation it represents, independently of any basis. The determinant is simply the signed volume change of the volume element.
Now we can see what a zero determinant means: it means the transformed volume element is zero. This transformation has squished it down into a lower-dimensional subspace, which means the mapped basis vectors can't be linearly independent. That means there's a nontrivial linear combination of basis vectors which maps to 0. In other words, there must be some [itex]x[/itex] in this basis such that [itex]Ax = 0[/itex].