Are you or have you worked for companies that overestimate their scaling needs or think they have "lots of data" when in reality they really don't? How do you/did you handle the situation? Did you inform them or just go along with it?
Every other company does this. All the time they overestimate
Yes. Gather data (including costs, actual usage, projected growth, and usage spikes), and write up a concrete proposal with well-thought-out alternatives (including keeping the status quo). Include engineering time when comparing costs.
Make it a discussion of tradeoffs, and go in willing to compromise and commit.