Lot easier to create a skeleton in the shape of a cylinder(cassion[1]) as you said, versus a sphere.
[1] http://www.thebluebook.com/inc/img/qp/75476/rebar-cages-for-...
The reinforcement cage in concrete weight coating used on subsea pipelines is relatively ordinary steel. The steel is protected from corrosion by the concrete itself and a system of sacrificial anodes that are easy to survey and replace. (I help design pipelines)
I guess that the steel is protected by something else because, as the pre-1970 builders ignored and as it dawned on everyone else 20 years ago, concrete is not waterproof.
Ancient Roman concrete is.
The engineering challenge is more comparable to submarine design. Nuclear submarines go down 500m and even the navy deep submergence rescue vehicle http://www.navy.mil/navydata/fact_display.asp?cid=4100&tid=5... is going 1.5km deep only with its tube design. Compare this with really, really deep dive designs like https://en.wikipedia.org/wiki/File:Trieste_nh96807.svg with a sphere observation gondola.
It's a pretty small-scale experiment, but everything has to start somewhere. The Deep Lake Cooling Project was prototyped at some point and that's well into production now, providing inexpensive cold water for chillers through the downtown core: http://www.canadianconsultingengineer.com/features/toronto-w...
The Hydrostor project is roughly located in the same area as the Deep Lake intake pipe comes ashore.
Even though one uses more concrete, it's cheaper because we already have the infrastructure in place to produce and lay long concrete pipes.
In both cases he also recommended using such energy storage systems as anchors to hold offshore wind turbines in place, because the concrete is more than heavy enough to do so.
[0]https://www.naefrontiers.org/File.aspx?id=48498