Except you can't just invest it all into renewables and expect to have a stable power grid, the equation isn't quite as simple as "avg_power / $ capex".
Last analysis I saw of this argued that the amount of "over-provisioning" needed to cover baseline demand with renewables made them prohibitively expensive as the sole solution. Peaker and storage solutions aim to solve that, but you need to build that capacity in addition to renewables.
Take a look at the graphs at the back. It shows a 99.99% reliable grid made up of an optimal mix of solar & wind, 3x over-provisioned along with 3 hours of storage and a continental grid.
We don't need 3x over-provisioning because the paper doesn't include hydro or existing nuclear. And 3 hours of storage is a lot of batteries, but it's a lot cheaper than building new nuclear.
Renewables are nowhere near to running up against the need to buy more grid interconnection, storage, synfuel technology, or other ways of using surplus power and providing power from storage or from other regions.
Until renewables displace a lot more fossil fuel power, those costs are not a consideration. That might happen within a decade in places like New Zealand.
