A pressure cooker clearly can withstand vacuum and float in water, which also is gas in liquid form. Soo then floating vessel with a difference of pressures of inside and outside is possible, in some conditions, then what about the generalization of that.
problem isn't simple if we approach the problem more seriously. Constructions working against compressive forces are a bit more complex subject and is as an example the serious field of research in NASA - as it is directly connected to the mass of a rocket and all that - they have a lab of experimental researches how exactly basically a thin-walled cylinder crumbles under a load.
The necessity of that lays in a difference of theoretical ideal construction and how much influence on a result imperfection of construction and materials have on the situation. The same applies to bridges and house building - finite element analysis goes hand to hand with experimental testing, and the safety factor is quite high, not only because it has to last long, but also for that negation of imperfection influences.
one of the places to start with theoretical parts is Euler's critical load and material science in general. I won't, not competent enough, I'ma doomer and it easier for me to look at Hydraulic press channel and they's not so long ago established dive 3km deep setup v.onemillion Shrinking Styrofoam Cups with Deep Sea Chamber and alike.
in general attempt to google how much a sphere will hold of outside pressure leads to strange places u never been and does not bring clear answers, even if it is a typical modeling task for the mechanics of materials field of science, and not much luck for those who aren't familiar with all that.
All that said, all below is not perfect, and no guarantees.
a little bit of trivia and water again
A submarine is a relatively big underwater vessel, which is capable to dive at significant pressures like 60 bar down, or ones of WWII like 20 bar typical.
And steering in direction of venus - density of liquid carbon dioxide is
1101 kg/m3 (liquid at saturation −37 °C (−35 °F))
Critical point (T, P) - 31.1 °C (304.2 K), 7.38 MPa (72.8 atm)
So a modern submarine potentially can float in liquid CO2, in all 1-74 bar pressure, in conditions when CO2 is liquid. The density of CO2 will change with the rising of temperature, but no so essential for us, a change of 10-20 percent or something - a difference in a range of ballast tanks.
let's start our speculations
The compressive strength of ductile materials such as mild steel used for most structural purposes is around 250 MPa.
that is the stuff nails made of - not the best, bends easily and all that, but we take that number as a measuring stick.
unfortunately lost my gnuplot skills so, let's jump straight to how it looks like for 100 bar.
Sphere 1m radius, crushing force will be 31MN, so crossection has to be 0.1256m2, hence wall thickness 0.02m, and resulting mass of our sphere around 1885kg
The volume of that sphere is 4.2 m3.
So the thing will float in something of a density of 450 kg per cubic meter.
- However, if we apply safety factor for all imperfections and all that, it may sink easily, but to not mess with it we took not the best material, even among the steels - one negates another and we can assume to have enough of safety factor here.
So half of the density of water, not bad, and considering real constructions which did dive at Mariana Trench, we are not that far off with all that.
how all that scales up/down?
pretty much linearly - crushing force scales proportionately to the square of linear sizes, so as resistance is proportional to the surface of cross-section, the strength of materials, mass proportional to cross-section aka wall thickness multiplied by surface area aka volume as the result, lifting capacity proportional to volume.
so no cheat codes like square-cube law. A variety of sizes is big enough before there will be some nonlinear effects, and they are, and 10m radius and we probably have to account for them, 10-20 percent range with steel, less for titanium.
let's plunge into Venus. Venus, my love, I come ...
what is the density of those 93 bars of pressure, near-surface of it?
wiki>>The density of the air at the surface is 67 kg/m3, which is 6.5% that of liquid water on Earth.
Soo, it seems we miss our target by 6.7 times. Density isn't high because of the high(relatively, 740K) temperature there, so we have the additional challenge of the steel to lose structural strength at the temperatures, but not a huge deal, the second order of magnitude effect, as there are steels which work well enough with higher temperatures.
unfortunately, there is no positive, for us in the case, square cube laws, so we can't outnumber the situation by changing the size. (maybe wrong, but do not see any, but can be mistaken)
So the variables to change are mostly density and strength of material with that density, composites, more sophisticated ways to squeeze their all from available materials.
But perspectives do not look that promising if we do not bring the new game in town like carbon something. it much easier to have carbon monoxide as lifting gas for those purposes. And if we cross the threshold with conventional materials then just barely - so imperfection of our original assumptions may be detrimental for the answer, as so the importance of experimental data can't be underestimated here.