The goal is to find the optimum values of the diameter and the thickness of the circular cross-section of a cantilevered beam in bending, with the lowest weight (cross-sectional area). The main criterion is to make sure that the internal stress does not exceed a limit anywhere in the cross-section. The length of the beam has already been chosen. The approach so far has been to find the internal loading of the beam from the external loads, then use the following formula. I am using the thin-walled assumption for the tubular cross-section.
$\sigma = \frac{M_{y}}{I_{yy}} x$
However, it appears that I can increase the radius indefinitely, with the same area, and always achieve a better cross-section.
$\sigma = \frac{M_{y}}{\pi R^{3} t} x = \frac{2 M_{y}}{(2 \pi R^{} t) (R^{2})} x = \frac{2 M_{y}}{(const) (R^{2})} x$
Is the limit solely based on manufacturability, or is there something else I should consider? Also, if it is based on manufacturability, could you give some advice on what the limits would be for composites or metals? I am currently considering either an aluminum alloy or carbon fiber. This is for a beam for an eVTOL vehicle that is connected to a propeller, so if you have comments relating to my approach let me know. I am currently considering only the vertical force from the propeller since I think it would be the critical load.