It's also worth mentioning that, while I'm not anti-nuclear on principle, the economic return on nuclear projects ranges somewhere between "multiple decades" and "never" - and there's a large empty gap on the timescale of a decade between spending most of that money and starting to receive dividends. And you'd better be running it 100%

At least with solar and wind the buildout takes a few weeks or months, and you can start collecting even with a partial buildout.

I'm afraid this bit - ROI time / profitability - is what will kill practical applicability of fusion power. There's already tens if not hundreds of billions in investments and decades of research, and it will take that again to turn it into a commercial endeavour, if ever.

That said, nuclear is great for baseline power production, and even with renewables generating the brunt of electricity, you still need a baseline and a quickly scaling backup (gas generators). Battery parks help too for those, but they have limited capacity of course.

This is a frequent argument against nuclear, that it’s too expensive(and takes forever to build).

But this wasn’t always the case, and so I wonder if it’s really such a strong argument.

In Sweden most of our plants were built fast and have been an enormous success.

The speed was partially enabled by an economy of scale that will not be possible to reach again like it was in the 50s / 60s / 70s.

Why wouldn’t that be possible?

Mainly because the government was backing the investment on national security grounds because of the cold war. Every variety of nuclear investment was through the roof

There was also massive demand for electricity as it was being extended to rural areas and as electric household appliances became common.

Until the AI craze there was no long term demand for that much additional electricity, and who knows if that will hold.