I think the reason they don't go the Thorium route is that Uranium is just so abundant -- and so incredibly energy dense -- there's not a hugely compelling reason to switch to Thorium.
They require an extra separation step (removing newly bred material from spent fuel, or removing fission products from molten salt). This is more expensive than just doing nothing to the spent fuel.
In the past, the idea was that nuclear would be cheap, but would run into uranium supply constraints, so breeding would save money. But that's not how it turned out. Nuclear was expensive not because of fuel, but because of the cost of the power plants. Uranium prices remain low. Also, the move to gas centrifuges reduced the energy consumed in uranium enrichment by a factor of 50.
My understanding is that uranium supplies remain constrained --- fewer than two decades if supplying 100% of total global generation, say. Price doesn't tell you much about total resource stock.[1] That's based on terrestrial sources. Seawater U separation in theory would extend resources considerably, but remains unproven at scale.
Thorium, other disadvantages notwithstanding, is at least more plentiful.
I'll note I'm not generally a fan of nuclear, though don't rule out any contributory role.
________________________________
Notes:
1. I'd argue generally that nonrenewable natural resource pricing theory, dating to Ricardo, but especially Hotelling, is entirely flawed. Much of it under suspicious circumstances.
Yes, but the "supplying 100% of global generation" doesn't have any bearing on current reality.
Anyway, this argument for thorium isn't something customers would care much about. It's basically "thorium would not suck as much as uranium-fueled burner reactors do if uranium gets much more expensive" rather than "nuclear power is more attractive now if we use thorium". The customer response to "nuclear as currently implemented fails badly if uranium runs out" is going to be "use something other than nuclear".
At some point we've got to address the question of how much energy is supplied to how many people and for how long.
Population is expected to rise for at least another 30-80 years, to between 9 and 12 billions by most estimates. These may not see US levels of energy access, but most authors project per capita energy wealth roughly comparable to present day European levels, largely as electricity. This represents multiples of present generating capacity.[1]
And energy represented by virtually any nonrenewable stock, including most fissionanbles, is finite. That's before allowing for technical limitations, concerns, wastes, risks, or other impacts.
When the U.S. was first transitioning from wood to coal, roughly 1860--1880, then-known reserves were calculated as sufficient for at least one million years at then-present rates of consumption.[2] The problem, of course, is that rates of consumption increased somewhat, by a greater rate than those of new coal discoveries. I can remember in the 1970s National Geographic adverts assuring readers that America's coal reserves were good for another 1,000 years, already a thousandfold reduction from 100 years prior. Today official estimates tend to run 200--300 years, though pessimistic ones suggest scarcely a century.[3] That's roughly 10,000 times sorter than initially anticipated, thanks largely to the Jevons Paradox: low-cost goods and increased efficience stimulate demand.
And all this before acknowledging that we simple cannot burn much more of the stuff.
So, no, I don't buy that "supplying 100% of global generation doesn't have any bearing on current reality.", as even a small fraction of a growing number, most especially an exponentially growing one, remains a large number.
________________________________
Notes:
1. 1991 per capita use, 287.8 GJ US, 75.3 GJ world, 123.6 GJ Europe. bp Statistical Review of World Energy 2020, p. 11. At 123 GJ * 12 billions souls, total global energy demand would be 1,476 EJ, vs. 584 EJ consumed 2019, a 250% increase. Total 2019 electric generation was 27,005 TWh, or 97.2 EJ, 16.6% of total global energy consumption. That works out to 3.46 MWh/capita, or about 395 W continuous per person.
3. BP's 2020 report gives an R/P ratio, reseves vs. production, equivalent to years supply at present consumption, of 390 years. This is an increase, though almost entirely due to reduced extraction, down from 22.27 EJ in 2011 to 14.30 in 2019. Consumption has fallen by more.
If we're talking about ultimate limits on Earth, solar is probably better than nuclear, just from direct thermal pollution. Solar causes a moderate increase in sunlight absorbed, but otherwise just moves solar energy around. If the albedo of the ground on which the solar modules were installed was less than the efficiency of the modules, there is no local heating, although the produced power gets degraded to heat eventually elsewhere as it is used. For every kWh of power produced, nuclear adds 3 kWh of heat to the biosphere (1 from the power produced, 2 from the waste heat of the reactor.)
(This is really a reflection of the difference between primary energy, which today is largely thermal, and delivered energy, which is largely work or chemical. The conversion to renewable energy will greatly reduce the importance of thermal energy conversion, and will not require a 1-1 replacement of today's primary energy use.)
It's not at all clear energy use will grow that much more. Lesser developed countries will use more, but in advanced countries energy use has plateaued. We are currently very far away from limits on solar energy imposed by shortage of sunlight. The Earth is hit by 100,00 TW of sunlight; global primary energy use is 20 TW.
If we're talking about limits OFF the Earth, solar is vastly more abundant than uranium (or, for that matter, artificial fusion, since the Sun fuses starting with ordinary hydrogen, not comparatively rare isotopes/elements like deuterium, lithium, or boron.)
The usefully convertable fraction of solar on Earth may be far closer to present or anticipated energy demands than is commonly thought. Panel efficiency, spacing factor, lifetime, capacity factor, storage requirements, essential fuel-based needs (marine shipping, powered flight, mobile power, remote reserve generation & thermal), process energy (steel coking, Haber-Bosch, etc.) leave some large holes and very uncomfortable margins remaining.
The alternatives to solar are either secondary or tertiary options (biofuels, wind, wave) and hence, more limited, or comparatively finite (geothermal, possibly our best non-solar option, tidal).
I do largely suspect that humanity's future will be principally solar powered. The question is largely of how much energy and in what forms it will be available. And aagain, demographic trends and expectations shade strongly against pleasant transition.
I find Vaclav Smil's and the late David MacKay's works quite illuminating.
