What is the optimal size for a controlled nuclear fusion reactor?

We have research currently from the nanotechnological level to, at least, the ITER Tokamak at 840 cubic meters, none of which seems to work very well, at all. So, what exactly is the optimal size for a controlled nuclear fusion reactor? Is anyone even really thinking about this? The problem with nanotechnology is that it never really works very well at all, for anything. The problem with the Tokamaks is, despite violent protests to the contrary, they're probably too big to be really safe, it they actually worked, at all. How about a Tokamak design with a very small reactor -- a cubic centimeter, say -- but not nearly at the nanotechnology level. So, we could have much higher plasma densities than would be practical in a conventional Tokamak, because the entire reactor could be embedded in reinforced steel and concrete to a considerable depth, without the slightest danger of anything more than some melted steel and concrete if the electromagnetic containment fails. Currently, at 2.6 atmospheres, the plasma density in Tokamaks is moderately dangerous, but not excessively so. At 100 atmospheres, for example, it might be very dangerous, but, it is also might work a bit better. It seems to me, that this smaller design would allow for greater options in design, that might be more effective. Particularly, much higher plasma density.