The $L_p$ norm $||A||_p$ is defined as
$$\sup_{x \neq 0} \frac{||Ax||_p}{||x||_p} = \max_{||x||_p = 1} ||Ax||_p \tag{1}$$
I'm not quite getting why the LHS uses $\sup$ but the RHS uses $\max$. I know how both are defined but somehow I don't get it. The maximum requires the solution to be contained in the candidate set while the supremum does not. Thus, why does (1) hold?
In finite dimensional space (which is the case here) the unit sphere $\{x\in\Bbb R^n\mid ||x||_p=1\}$ is compact then the supremum of the continuous function $x\mapsto ||Ax||_p$ is attained.