The issue is the speed of light is so big. So, you need a LOT of energy in order to source gravity comparably to just a little bit of mass (essentially, because E=mc^2 -- or perhaps phrased more clearly in this case, m = E/c^2).

Even Newtonian Gravity can bend light classically, by Galilean relativity. But, weirdly, it essentially hinges on the fact that the the mass m of the photon cancels from m*a = GmM/r^2. Of course, that is really the equivalence principle---the acceleration of all things under gravity's influence is the same. Whether or not light is a source of Newtonian gravity... I'm not sure. It's a tricky question because m=0. I want to say no because the "equal and opposite" forces should both be 0, even though one of them effects an acceleration on the other. I should emphasize that I'm not sure!

In Einsteinian gravity, the paths of photons (and indeed all things) are bent because spacetime itself is curved. Classical electromagnetic static fields and waves certainly have an energy density that can source gravity, and individual photons do too. But their energy is on the order of hbar. So you're talking a source of gravity like hbar/c^2. THIS IS REALLY TINY unless the photon's frequency is ENORMOUS.