You can't talk about statistical significance without actually having data.

It sounds more like what you're interested in is doing a sample size calculation. Typically for these you specify a specific alternative hypothesis and a power you would like to achieve. Once you have that information you can solve for the sample size required.