Of course that "limitless" part is pure marketing.

We agree on that, but I think it is worth looking at how limited fusion might be.

Even if operating costs are virtually zero, if the capital cost is very high, then the cost per kW could still be more than a fission plant.

A fusion reactor has similar engineering limits to a fission plant, so a fusion unit might produce 1500MW with a lifespan of 40 years. They are still going to need a hefty injection of capital to replace a significant portion of energy generation.

If the capital cost is too high, they may not even be economic with wrt to other forms of generation. In that case, only rich countries with a healthy economy could muster the capital to build them.

We could end up in a situation where we have a theoretical "unlimited" power source, but be unable to afford to build power plants to exploit it.

If anyone has a handle on the numbers, what would be the maximum capital cost of a fusion plant before it became uneconomic to build?

This is entirely valid point. It has always been the assumption that with technology development costs will drop. But what if the nature of the technology itself does not allow it?

In this case it is entirely possible this to happen, largely because of the material costs for building such a huge plant. I am not familiar what are the limiting factors for fission reactor and if they automatically apply for fusion but your claim that the size will be similar looks reasonable. In this case it does not look good IMO.

Basically the jury is still out. It may turn out that the lower cost of fuel, the simpler safety eqipment and lack of radioactive waste etc. will make up for the other increased costs. But it also may not, we are still in the R&D stage and it's too early to tell. This I guess is one of the goals of building ITER and I am 100% sure that any commercial plant will be built only after it is understood that it will be profitable. 10 bln. is not such a big price tag to find out.