I think it's fairly unlikely that someone is going to come up with a revolutionary new algorithm that would make the actual tracing of rays dramatically faster.

I think it's fairly unlikely that someone is going to come up with a revolutionary new algorithm that would make the actual tracing of rays dramatically faster. People have been researching better acceleration structures and optimization tricks for decades at this point. There is probably a decent amount of scope for incremental improvements via hardware acceleration though. There is also quite a bit of scope for figuring out clever ways to get the maximum value out of a given number of rays cast. This includes high level engine architecture and lower level improvements to things like denoising algorithms and smarter importance sampling. I think there's still a lot of room to better exploit temporal coherence for realtime ray tracing.