Use of nuclear could preclude the power issue.
Anyone ran the numbers on how long uranium would last running sun-less agriculture for a few billion people?
When fusion arrives in the next 40 years: "Lithium from sea water would last 60 million years, however, and a more complicated fusion process using only deuterium would have fuel for 150 billion years." [calculated at 1995 global power output]
One thing people don't get is we only need a tiny fraction of net positive output due to the insanely large amount of fuel available to us. 2-3% positive output is all we're talking about and it'll be able to economically replace most other energy sources (though it will take time to ramp up of course.)
Basically it's just a matter of time and continued advances in material sciences. Purity of materials and perfection of manufacturing are the primary barriers. Operational consistency a secondary one (and continued improvement in microcontroller tech and data analysis/AI/etc. will help here.)
One example, in focus fusion, the current progress has definitely been incremental but basically every step of the way the progress has been hindered by exactly the same problem: the materials and apparatus have to be nearly perfect to increase plasma duration sufficiently; to avoid contamination of the self-sustaining plasmoid. And indeed, the progress on that end has been continuous. Once it's proven it will be a matter of perfecting the manufacturing process and automations/etc. to ensure that degree of material quality and device engineering and voila we'll have cheap scalable power.
Based on what I've been watching for the last 15 years, I think within 40 years is quite reasonable. In fact, I think it would not be a surprised to see consistent 0.5% net positive or better in the next 5-10 years, whether it is focus fusion or tokamak or whatever.
TL;DR: Fusion is a 100% sound energy generation avenue that is simply waiting for its "welding" moment. By which I refer to the point where gas production became suddenly economically viable due to the invention of welding (which was motivated at least partly by the desire to improve gasoline production.)
Right now human civilization is running off something like 18TW of energy.
If you add in 2000 kilocalories per day for 7.8 billion people, you add another 750 GW; if you add in a factor of 10 for biological & technological inefficiencies on actually turning electricity into edible calories, you're up to 7.5TW.
As long as you don't throw another factor of 10 on top of that -- imagine growing all food indoors, and then feeding it to cows, before subsisting entirely on beef -- you're probably looking at an amount of energy comparable to that used by humans currently.
Obviously it doesn't make a lot of sense to just switch over to indoor agriculture when there's free sunlight outside, but if we were to do so, the energy demands are probably on a scale comparable to what we're working with already.
So you're probably only halving or so the runway of available uranium.
What about recycling for wind/solar?
Edit: Got the dates wrong.
[0] https://en.m.wikipedia.org/wiki/World_energy_consumption
Sure, this can be used used land not suitable for agriculture. I only see this really being helpful in cities that have old industrial buildings that are sitting vacant (reuse existing infrastructure and reduce transportation).