Based on some computational experiments, it seems that, if $x_{0}, x_{1}, \dotsc, x_{n} \in \mathbb{Z}_{\ge 1}$ are such that \begin{equation*} \left(\sum_{i=1}^{n} x_{i}\right)^{2} = x_{0}\prod_{i=1}^{n}x_{i}, \end{equation*} then $x_{0} \le n^{2}$.
Is this bound known to be true? And (to ask a less precise question) whether or not this bound is true, where should one get started to learn methods for proving bounds like this, assuming no background beyond elementary number theory?
Here is a solution. Thank you to Will Jagy for suggesting Vieta jumping. I had been trying Vieta jumping, but I guess I needed a bit more encouragement to push it through.
Let $x_{0} \in \mathbb{Z}_{\ge 1}$ be such that \begin{equation} \left(\sum_{i=1}^{n} x_{i}\right)^{2} = x_{0}\prod_{i=1}^{n}x_{i} \label{eq:1} \end{equation} for some positive integers $x_{1} \le \dotsb \le x_{n}$. Using Vieta jumping, we may assume, without loss of generality, that $x_{1} \le \dotsb \le x_{n} \le \sum_{i=1}^{n-1} x_{i}$. For, Equation \ref{eq:1} is equivalent to \begin{equation*} x_{n}^{2} - \left(\prod_{i=0}^{n-1}x_{i} - 2 \sum_{i=1}^{n-1}x_{i}\right) x_{n} + \left(\sum_{i=1}^{n-1}x_{i}\right)^{2} = 0. \end{equation*} Letting $x_{n}'$ be the other root of the polynomial $t^{2} - \left(\prod_{i=0}^{n-1}x_{i} - 2 \sum_{i=1}^{n-1}x_{i}\right) t + \left(\sum_{i=1}^{n-1}x_{i}\right)^{2} \in \mathbb{Z}[t]$, we have that $x_{n}x_{n}' = \left(\sum_{i=1}^{n-1}x_{i}\right)^{2}$. If $x_{n}$ is already the smaller of these roots, then we are done. Otherwise, we have a solution $(x_{0}, x_{1}, \dotsc, x_{n-1}, x_{n}')$ to Equation \ref{eq:1} with $x_{n}' < \sum_{i=1}^{n-1}x_{i} < x_{n}$. Reordering $x_{1}, \dotsc, x_{n-1}, x_{n}'$ in increasing order, and then reapplying this reduction step a finite number of times, we eventually arrive at a solution of the desired form.
Now, given a solution $(x_{0}, x_{1}, \dotsc, x_{n}) \in \mathbb{Z}_{\ge 1}^{n-1}$ to Equation \ref{eq:1} with $x_{1} \le \dotsb \le x_{n} \le \sum_{i=1}^{n-1} x_{i}$, we solve for $x_{0}$ and find that \begin{align*} x_{0} &=% \frac{\left(\sum_{i=1}^{n} x_{i}\right)^{2}}{\prod_{i=1}^{n}x_{i}} \le % 4\frac{\left(\sum_{i=1}^{n-1} x_{i}\right)^{2}}{\prod_{i=1}^{n}x_{i}} \le % 4\cdot% \frac{\sum_{i=1}^{n-1} x_{i}}{x_{n}} \cdot \frac{\sum_{i=1}^{n-1} x_{i}}{\prod_{i=1}^{n-1}x_{i}} \\ &= % 4\cdot% \sum_{i=1}^{n-1} \frac{x_{i}}{x_{n}} \cdot \sum_{i=1}^{n-1} \frac{1}{\prod_{\substack{1 \le j \le n-1 \\ j \ne i}} x_{j}} \le % 4\cdot (n-1)\cdot \frac{n-1}{\prod_{j=1}^{n-2} x_{j}}. \end{align*} Thus, if at least two of the numbers $x_{1}, \dotsc, x_{n-2}$ are $\ge 2$, or at least one of the numbers $x_{1}, \dotsc, x_{n-2}$ is $\ge 4$, we have that $x_{0} \le (n-1)^{2}$, and we are done. Otherwise, $(x_{1}, \dotsc, x_{n})$ has either the form $(1, \dotsc, 1, 1, x_{n-1}, x_{n})$ with $x_{n-1} \le x_{n} \le n-2 + x_{n-1}$, or the form $(1, \dotsc, 1, 2, x_{n-1}, x_{n})$ with $2 \le x_{n-1} \le x_{n} \le n-1 + x_{n-1}$, or the form $(1, \dotsc, 1, 3, x_{n-1}, x_{n})$ with $3 \le x_{n-1} \le x_{n} \le n + x_{n-1}$.
In each of these cases, one can work out directly that $x_{0} \le n^{2}$. ETA: For example, if $(x_{1}, \dotsc, x_{n})$ has the form $(1, \dotsc, 1, 3, x_{n-1}, x_{n})$ with $3 \le x_{n-1} \le x_{n} \le n + x_{n-1}$, then \begin{align*} x_{0} &=% \frac{(n+x_{n-1}+x_{n})^{2}}{3x_{n-1}x_{n}} \\ &=% \frac{n^{2}}{3x_{n-1}x_{n}} + \frac{2}{3}\left(\frac{1}{x_{n-1}} + \frac{1}{x_{n}}\right)n + \frac{1}{3}\left(\frac{x_{n-1}}{x_{n}} + 2 + \frac{x_{n}}{x_{n-1}}\right) \\ &\le% \frac{n^{2}}{27} + \frac{2}{3} \cdot \frac{2}{3} \cdot n + \frac{1}{3}\left(1 + 2 + \frac{n + x_{n-1}}{x_{n-1}}\right) \\ &\le% \frac{n^{2}}{27} + \frac{4}{9} n + \frac{1}{3}\left(3 + \frac{n}{3} + 1\right) \\ &\le% \frac{n^{2}}{27} + \frac{5}{9} n + \frac{4}{3} \\ &\le% n^{2} \quad \text{for $n \ge 2$}. \end{align*}
Thank you again to Will Jagy for pointing out a missed case.