I was told by faculty with joint appointments in matsci and nuclear engineering that it is a major issue and that while they've developed some alloys that look reasonable in the recent past it's still academic research unproven in scale up.
Holding molten salts is really hard... they corrode the heck out of even high end commodity stainless steels (this I know from professional work). Protective coatings help for sure, but they crack over time, especially with any thermal cycling. Inconel and similar alloys maybe, but it's really not straightforward to block grain boundary corrosion! Add to that neutron embrittlement...
I've not kept super up to date, but it's hard to imagine that these alloys went from academically unproven to industrially certified in the last seven years.
Specifically you should read the findings of the MSRE (Molten Salt Reactor Experiment) at ORNL.
Wikipedia has a relatively laymans explanation of the state of the Hastelloy-N when the experiment concluded.[1]
It's worth mentioning Hastelloy-N isn't that expensive all things considered. It would probably come out to a small percentage of the reactors total cost factoring in R&D expenditure etc.
[1] https://en.wikipedia.org/wiki/Molten-Salt_Reactor_Experiment...
+ Ethylene-oxide gas + moisture + heat -> plying rough trade on supposedly compatible materials.
Not my problem but a customer I worked with had a gamma ray sterilization unit. There were some limit switches which they needed to replace after ten years. Replacements, same part number, manufacture swore up and down they were identical, would crack after three to six months.
At least two startups are going that route. ThorCon's design is a plant with reactor cores as replaceable "cans," each lasting four years in production plus four years of cool-down before shipping to a reprocessing facility. Terrestrial Energy has sealed 7-year units.
As for water, the working fluid of the heat exchanger and generator can be water, though Helium or CO2 would make more sense.
Additionally, renewables are absolutely not ready to take over base load power generation. They are an incredibly flawed system in their current form, and incredibly expensive (in multiple ways). They can be improved but it'll be a long, long way before we can rely on solar and wind as the core (rather than periphery) of our power generation infrastructure.