Also you need to lift your fuel and you might get into a rocket equation type problem. Fuel is less dense than water so one might imagine it contributes a little to buoyancy but aloft, you're paying to lift it along with the cargo.

Fuel means range so that all might get back to the mission: ferry or transoceanic.

Airplanes also have to lift their own fuel, yet are not, in practice, hostage to the rocket equation (though early, grossly inefficient jets may have experienced something like it.) Three things that make the rocket equation a problem for orbital launches are, firstly, you have to take along the oxidizer as well; secondly, you have to do much more lifting (and, considerably more significantlty, accelerating) than an airplane; and thirdly, without dynamic lift, all the lifting and accelerating has to be done directly by engine thrust.

Also, while fuel oil is buoyant, it still has mass and weight, and the fuel in a ship contributes its own weight's worth of displacement. Displacement, in turn, causes wave resistance.

The larger a ship gets, the less significant, proportionally, both wave and friction drag become, reducing the incentive for foiling. It seems significant that the YC company mentioned elsewhere in these comments (Boundary Layer Technologies) positions itself as competing with air freight rather than shipping (and, as flying in the thin air of the tropopause is pretty efficient compared with generating lift in water, I would guess part of their calculation involves building much larger-capacity vessels than current and projected air freighters.)