When the US was first making the switch from wood to coal as a principle fuel, in the 1880s, the argument in contemporaneous accounts was that a small fraction of the probable reserves of coal would serve all energy needs for millions of years, given that perennial caveat, at present rates of consumption.
What happened, of course, was that rates of consumption increased somewhat. Coal went from fueling a small number of structural heating, locomotion, and smelting uses to vastly expanded railroad, steelmaking, and most especially, electrical generation uses. Trainloads of coal arrive daily at power plants, still. Others are located at mineheads themselves as electrons transport more readily than lumps of fossilised tree. The Jevons paradox is a mighty bastard. By present estimates, again, at present values of consumption, coal reserves are estimated at 100-300 years.
At present rates of electrical energy generation, existing terrestrial (non-oceanic) uranium would suffice for fewer than two decades worth of energy use. When you factor in even probable growth due to:
- Extending Western standards of living to another five billions of the world's population.
- Expansion of total population from 7.5 billions to 11-12 billions.
- Continued "normal" economic growth at 2-3% per annum, that is, doubling every 25-30 years, roughly.
... the total energy requirements increase tremendously.
(Note that renewables also strain to keep up with such growth -- my sense is that the fact that incident solar power on Earth of some 7,000x present human energy consumption actually represents a dangerously narrow margin of safety.)
I'd worked out at one point the remaining reserves given some rate of projected annual growth, a fairly simple application of exponential math. The numbers are exceedingly sobering.
TL;DR: If there is some putatively nonrenewable fuel source whose reserves would extend to some point in time beyond the viability of some specified landmark (h. sapiens as a viable species, C3 photosynthesis on Earth, etc., variously a few million to about 800 million years), then the resource is not meaningfully constrained, and we can make reasonable use of it.
But given any notion of continued exponential growth, which is to say, a constant percentage increase in economic activity, given all available history, if not some putative but very-much-unproved virtualisation-of-activity hypotheses, that's a high bar to meet.