[guided]We prove nonvanishing at a single point $p\in M$, because a differential form is nowhere vanishing precisely when its value at every point is nonzero. If $n=0$, then $M$ has dimension $0$ and the convention for exterior powers gives $\omega^0/0!=1$, the constant $0$-form, which is nonzero at every point. We now assume $n\ge 1$.
The vector space $T_pM$ has dimension $2n$, and $\omega_p$ is a nondegenerate alternating [bilinear form](/page/Bilinear%20Form) on it. By the [Symplectic Basis Theorem][citetheorem:10037], we may choose a basis
\begin{align*}
(e_1,\dots,e_n,f_1,\dots,f_n)
\end{align*}
of $T_pM$ satisfying the symplectic basis relations
\begin{align*}
\omega_p(e_i,e_j)=0
\end{align*}
for all $1\le i,j\le n$,
\begin{align*}
\omega_p(f_i,f_j)=0
\end{align*}
for all $1\le i,j\le n$, and
\begin{align*}
\omega_p(e_i,f_j)=\delta_{ij}
\end{align*}
for all $1\le i,j\le n$.
Let
\begin{align*}
(e_1^*,\dots,e_n^*,f_1^*,\dots,f_n^*)
\end{align*}
denote the dual basis of $T_p^*M$. In this dual basis, the defining pairings above say exactly that
\begin{align*}
\omega_p=\sum_{i=1}^n e_i^*\wedge f_i^*.
\end{align*}
The reason this normal form is useful is that the top exterior power can now be computed by pure algebra.
For each $1\le i\le n$, define
\begin{align*}
\alpha_i=e_i^*\wedge f_i^*\in\Lambda^2T_p^*M.
\end{align*}
Then $\omega_p=\alpha_1+\cdots+\alpha_n$. Since every $\alpha_i$ has degree $2$, swapping two such forms gives the sign $(-1)^{2\cdot 2}=1$, so
\begin{align*}
\alpha_i\wedge\alpha_j=\alpha_j\wedge\alpha_i.
\end{align*}
Also,
\begin{align*}
\alpha_i\wedge\alpha_i=(e_i^*\wedge f_i^*)\wedge(e_i^*\wedge f_i^*)=0,
\end{align*}
because an alternating wedge product vanishes whenever a covector occurs twice.
Now expand
\begin{align*}
\omega_p^n=(\alpha_1+\cdots+\alpha_n)^n.
\end{align*}
Any term in this expansion with a repeated index contains $\alpha_i\wedge\alpha_i$ for some $i$, after commuting the degree-$2$ factors, and therefore vanishes. The only nonzero terms are those in which every one of $\alpha_1,\dots,\alpha_n$ appears exactly once. These terms are indexed by the $n!$ permutations of $\{1,\dots,n\}$. Since the $\alpha_i$ commute, each of those $n!$ terms equals $\alpha_1\wedge\cdots\wedge\alpha_n$. Therefore
\begin{align*}
\omega_p^n=n!\,\alpha_1\wedge\cdots\wedge\alpha_n.
\end{align*}
Dividing by $n!$ gives
\begin{align*}
\frac{1}{n!}\omega_p^n=\alpha_1\wedge\cdots\wedge\alpha_n.
\end{align*}
Using the definition of the $\alpha_i$, this becomes
\begin{align*}
\frac{1}{n!}\omega_p^n=e_1^*\wedge f_1^*\wedge\cdots\wedge e_n^*\wedge f_n^*.
\end{align*}
This is the exterior product of all covectors in a basis of $T_p^*M$, so it is a nonzero element of the one-dimensional vector space $\Lambda^{2n}T_p^*M$. Hence
\begin{align*}
\left(\frac{1}{n!}\omega^n\right)_p\ne 0.
\end{align*}[/guided]