Getting a million people to Mars, assuming Musk's price optimism is accurate, is somewhere around 10% of the upgrade costs Earth's electricity grids already need, or about half the material (not land, not planning, not political) cost of building a new circumglobal 1Ω TW-scale power grid from aluminium.

I would (if I had a say in the matter) go for the Moon over Mars or an orbital space station. An orbital space station (or collection of them) for a million people would be very hard, because we don't have a convenient small asteroid to mine for building materials and therefore can't ISRU. I'd pick the Moon because it's close enough that when the inevitable black swan catastrophe happens, it's only a few days travel so we'd have a chance of saving almost everyone's lives if all the water or food was lost to depressurisation/the outside environment/surprise mould.

Mars, to rephrase what I wrote before, is less hospitable than all the worst environmental disasters we can experience on Earth, even if those disasters were combined. Being able to colonise it means we can fix basically anything that goes wrong here.