One of the game changers are new superconductors. These are smaller, lighter and cheaper than anything we've had before. They're a very recent product of American and European colliders. Their use is common across many compact designs.
> Anything that will be employed in a design claiming bigger efficiencies has to be using less than a bare minimum used by them
Not necessarily. For example, a low-efficiency design that is simple to mass manufacture could be cheaper to build and operate than an expensive, efficient design.
Furthermore, for each element you list as being necessary there are designs which do without them. For example, Tri-Alpha uses direct energy capture (no steam power plant). Several companies eschew refueling in favor of capsulation. Large-scale forging is an odd thing to list when we're talking about minitiarrisation. Et cetera.
A small amount of research would show many arguments you present as established fact to be untrue or unfounded.
The other big savings for an MSR is a very large reduction in the size of the containment building. This is because there isn't water in there that would turn to large volumes of steam during an accident and that would have to be contained by a large pressure-bearing structure. Moltex claims their containment would be a factor of five smaller per unit of power output, compared to a LWR. This would be a huge cost advantage.
Nuclear engineering is by far not a barely-explored and novel technological domain, but an industry with near 70 year history with extremely high barriers to entry.
For you, it should be reasonable to believe that engineers and physicists with multiple postdocs, people of far greater intellectual achievements that you, me, and probably most of this website's demographic, were banging their heads non-stop for the last 70 years on the problem of making nuclear power stations economical. Nothing what came from years of their work indicate that anything but big and simple PWRs and BWRs makes sense economically.
It took decades to shave everything that can be shaved off the construction of gen 3 plus reactors to arrive at their current designs. Anything that will be employed in a design claiming bigger efficiencies has to be using less than a bare minimum used by them and that's not a lot:
- Nuclear safe superalloys - check
- Large scale forging - check
- Refueling infrastructure - check
- No mechanisation beyond control rods and pumps - check
- Advanced sensing - check
- Fuel assemblies - ridiculously cheap in relation to everything else
- On-line chemical filtration and corrosion control - check
- Steam power plant - check
Really, the only opening for lower costs are reduction in operation expenses - less scientists on site, less consumables, less or not manual labour intensive servicing, less frequent and cheaper refuelling. Besides refuelling costs, that's not a lot.