Earlier quoted context omitted.
There are those who wants to go larger, by building a submerged collider[1] in the area formerly known as the Gulf of Mexico which could do 500 TeV. Using established technology from offshore industry it's supposedly not as crazy nor costly as one might first imagine. There was a nice and fairly accessible talk[2] given at Perimeter Institute, which gave some background and went into this and the FCC. [1]: https://ar…
The link mentions 100 meters depth. Without being an expert that sounds deep enough to withstand hurricanes and such. However 100m depth? For a structure _that_ size, which I think might contain high vacuum, let alone systems that might need humans to service when they break? What is the feasibility of that problem?
Also mentioned is that the LHC cryostats, some nice illustrations here[1], are very nearly neutrally buoyant. Given these would be built similarly, they wouldn't need significant infrastructure to keep them in place.
So no humans would need to go down there, if something breaks they could just replace the broken segment(s) and fix the broken one(s) on the surface.
[1]: https://cas.web.cern.ch/sites/default/files/lectures/bilbao-...