Even if we assume for a moment that we got it working, the factors are: scale and embrittlement. Current thermonuclear technology, theoretically, breaks even, energywise, on reactors of certain scale. There is a reason to believe that once we build a reactor of a certain size, the energy output of a reaction will be bigger than the energy spent on jumpstarting it and magnetic containtainment. embrittlement of reactor structural materials is caused by radiation. After a while, reactor has to be rebuilt from totally new materials.

>After a while, reactor has to be rebuilt from totally new materials.

The closest parallel we have is nuclear reactors. Navy Nuclear submarines are probably the most prevalent safe small reactors (US submarine reactors not K19) and they have significant lifespans extending across multiple decades. (source: passed my navy nuclear engineer exam in a past career) The resulting embrittled can/containment does need to be handled safely, but this is a capital depreciation that can be modeled and planned for.

The MIT design essentially solves those two issues. Scale is smaller with stronger magnetic fields, and embrittlement is confined to an inner wall which is easily replaced once a year.

You are correct! It’s maddening that you had to make the correction, though. Commenters in here seem incurious about the new developments detailed in the source material.

There are fusion designs that heat water and there are some that can capture electricity directly. Who knows which will prevail but the the latter should be cheaper and more compact. Maybe so compact that it could be used on ships, trains, or spacecraft.

Nobody has a clue, because we haven't figured out how to build a reactor that generates net positive power at all yet. The cost of the fuel is trivial, but there's no telling how much it will cost to mass-produce and maintain stuff like superconducting magnets, high-grade vacuum systems, whatever we end up using to contain the massive neutron flux, etc.

Sounds like $1.8 per watt once its scaled:

https://youtu.be/KkpqA8yG9T4

Tritium. In addition to the points made by other replies, current fusion is (mostly) D-T, and we’re nowhere near breeding T in the reactor, so it has to be made in fission plants. Tritium is very expensive, and a significant risk for nuclear weapons proliferation.

This will probably end up pushing us to Lithium fusion pretty fast, the fuel is a lot more common and the temperature requirement is only a little hotter.

That said nuclear proliferation from tritium is a lot less of a risk than from uranium. Anyone can put together a nuclear event with sufficient enriched uranium, pinching tritium to fuse using a fission event, well it's a significant technical hurdle.