That size/scale looks rather expensive (judging that to be 80-90' high or about 30 meter based on the number of floors and the average human scale height of a building's floors or a ship's deck), but I think a reasonable case could be made for the dimensions I proposed above without too many investors.
Especially submerging past the mid-point. It occurred to me that if the mid-point is in the wave zone (+- the average wave height from the average calm waterline) there might be a significantly larger problem with scouring action affecting the sea bed under the structure.
Which means probably no day-lighting for any decks below the mean high tide + wave action point at least on the exterior of the torus.
And lack of fresh water is the flaw of most islands that are unoccupied right now in the world. Rather than sand ballast, I would go with a fresh water reservoir. The nice thing is the ballast to fill it could be towed to the site at the same time in a polymer bag and pumped in once situated.