The reason probably due to the catalyst or template used in the preparation.  Even if we change the shape of the template, the resulting materials are almost same.  If we carry out the synthesis through Hydrothermal method, the result are different.  A variety of size and shape are obtained.

Not all materials are spherical. Those that are spherical is because they start with nucleation and slowly grow to larger and larger. There are also materials that start as seed crystals and then grow from different facets. This results to rods, romps and other interesting geometric shapes.

I don't think so most of materials synthesized by Sol-gel  are spherical, however,  dimensionality is depend on deferent factors such as,  Temperature, Precursor, solvents, catalysis ..., and I am agreed  with Georgios Pyrgiotakis. 

spherical is preferred (except for very special cases where the growth in 1 directio is preferred compared to another one) because:

colloidal systems have interfaces, the system wants to minimize the interfacial energy; spheres have the smallest surface area for a given particle size, therefore, with the spherical form, the systems can minimize the interfacial energy (which anyway is far from thermodynamical equilibrium, we are dealing - in thermodynamical terms - with non-equilibrium processes and non-equilibrium systems)

Sol-gel method is preferred to reduce the size of the particle, gel formation reduces (tends to decrease the surface area) surface area, minimum surface area shape is spherical, so most of the particles are in spherical, but not all.