To me, this means that planetscale (and vitess) immediately become less useful for a large class of applications: existing apps that use an RDBMS that you want to scale. Because most existing apps will expect FKs. I haven't tested my employer's application (which has FKs) to see what will happen, but I doubt it will be smooth sailing.
But there's probably plenty of greenfield development where teams would prefer to stick with standards like SQL rather than a proprietary noSQL language, and that seems like a great fit for this solution.
CitusDB for example handles this by allowing FKeys between tables that share the same partition key. However it's also not -as- scalable as Vitess or as georeplication friendly but for many uses that need "more than a single box" is probably enough scalability.
And the reason why is basically “physics”. The speed of light is pretty damn slow all things considered.
Until we find ways to make all the servers in the cluster have latency measured in something like picoseconds, something is gonna have to give.