Also building enough battery capacity to offset the night is doable. UK annual consumption is around 300TWH, let's round to 1TWH a day.
The nighttime consumption is about half of the daytime's so only 33% needs to be saved.
So we need 350GWH of batteries. It will take a while, cost 35 Billion assuming 100 usd for 1 kwh. Or 2 billion a year or 0.6 cents per KWH of energy consumed.
Also, it's interesting that Iceland, with their abundance of geothermal power, don't have an interconnection with Europe. That's only 1000km as the crow flies and yet hasn't been considered economically feasible yet.
As for geothermal, Iceland has geothermal; yes. So do other places. You're talking as if using that as an infinite source of energy is both cheap, feasible, and scalable. It's probably none of those things. Iceland has about 300K people, the UK has around 66M, so we're talking about 220x the energy demand. And running cables across the Atlantic. And doing a lot of infrastructure development in Iceland at the cost of many billions in a market that has energy prices trending down because of other clean energy options. Perhaps that's the reason that it was never really considered as an option.
Regarding the quoted prices, I think for the sake of argument this person was deliberately using numbers that are on the super conservative side. E.g. buying battery at 100$/kwh would probably be a spectacularly bad deal considering prices are already dipping below that (and that's for car batteries). Also, for the sake of argument assuming everything happens right now at today's prices is arguably not how this would play out at all. But as he's argueing, even if you make such assumptions it's still doable. So, it's likely to end up being a good deal cheaper than projected there as this plays out over the next decades as prices for batteries continue to drop and we figure out how to do this at scale.
My point about Iceland is they already have huge amounts of geothermal power and they should be able to develop it and sell it back to Europe as green electricity. Currently 70% of their power consumption is used to smelt aluminium so it's not like they are constrained.
It's not useless, it just requires too much processing and transportation to be practical. The issue with water is that for what we need it, we really require enormous amounts of it. This is not the case with lithium, where required amounts are trivial in comparison. Just to make the point: a truck-load full of lithium is worth around $300,000. To desalinate the same truck-load worth of salt water, it costs about $20, that is, 4 orders of magnitude less. If you don't have to desalinate, but instead can just draw fresh water, it will be around $1-2 for a truck-load. Obviously, trucking in water is still too expensive, because we need lots and lots of water, truly insane amounts compared to anything else.
The sun is always shining somewhere doesn't work for day / night solar production. Transmission losses are of the order of 0.5% per 100km. By the time you transmit power 20,000km there is nothing left.