Land usage or energy density or amount of materials used are all red herrings.
I sometimes see mention "laws of physics" as a reasons renewables won't work, but nobody has ever, literally ever, been able to point to the law. Or run the numbers on why renewables would be be feasible for supplying all our power and even an order of magnitude more than we currently consume.
Since you say "laws of physics" prevent renewables from supplying out energy. I call bullshit.
Name the law.
So, I maintain that this is 100% BS, just an argument to authority. If there is an argument in hat book, do you understand enough to give a sketch? Or point to the principle?
Replace all energy and not just what is already electrified, and also boost global energy use to the per-capita rate of Qatar (I think the highest in the world at about 2.5 times the average of the USA), and also boost world population to 10 billion, you can still do this 30 times over with PV placed slightly worse than if it was randomly scattered.
There's a lot of land on this planet of ours.
That's why land use is a red herring.
In famously sunny washington state a current technology solar storage system with the same net capacity factor as US nuclear and the same number of days where >0 backup is needed as European nuclear for less than what Vogtle cost per watt. Add wind and this goes down dramatically. Allow the builder to sit on their thumbs for a decade before starting and they'd open before a new nuclear build while providing 100% capacity and a quarter of the cost using batteries made of abundant materials and solar panels made with the silver and copper that came out of existing uranium mines.
The average american has 200-400m^2 dedicated to parking their car. Worldwide average per capita primary energy is 2kW. Even at 20W/m^2 this provides enough. Even wind at 5W/m^2 can produce enough final energy in that much area (with the side benefit of not really affecting other uses).
Wind is much better than solar economically at high latitudes and is slightly anti-correlated with solar.
There are terawatts of tidal sites available in high latitude coastal areas and the LCOE is already lower than nuclear with a much better learning rate and only a few hundred MW of preexisting capacity.
What law of physics contradicts this?
The inkai uranium mine is one of the largest mines in the world. It spans about 760km^2 of desert. At that latitude current generation solar has about 16% capacity factor of 230W/m^2. There are much denser uranium reserves, but if land use being significantly low than solar is your overriding concern then there's only a few decades worth of reserves available for use in PWRs.
A solar project there would produce about 25GW net.
The uranium mined produces 50GW gross at typical burnup. Mining, transport, enrichment, and post burn lifecycle consume ~10-20% of this.
The exclusion zones for those reactors would be roughly another 170km^2 so the nuclear is only ~30% better.
Add wind turbines at 10W/m^2 and do it not in the arctic and the renewables win handily by net output. You can also use all of that land for other things as well without getting radon poisoning or being shot by armed guards. Modern multi megawatt wind turbines don't meaningfully use land as being able to see two or three turbines off in the distance doesn't really prevent you from doing most things.
As to variability, the people of france would like a word. Less facetiously you should consider cost to meet a minimum power target. With the $23k per kw of Vogtle, hitting higher reliability than any nuclear reactor is trivial -- you also get 10 or so kW of variable power.
https://medium.com/@alkidel/the-land-footprint-of-solar-and-...