It's because Google has lots of internal services (sorta like microservices) and that's how services communicate with their neighbors. In some cases, persistent data is in protobufs too.
The point is probably doing some kind of storage, calculation, or aggregation of data, but "moving protobufs" sounds funnier.
At the end of the day, all software engineers are moving, transforming and persisting bits.
Languages abstract bits to types. But, there is no universal langauge for storage, retreival, front-end rendering and middle-layer logic, so engineers have to fall back to sort of thinking in bits again as they interact with various systems.
Protos are Google's fairly successful attempt at raising the abstraction level to object / document so that programmers can store, retrieve, transform objects at all layers.
It's not really because of the monrepo. Think about how many jobs outside Google involve moving JSON data around (and manipulating and transforming it). At an Alphabet company, that work would likely use protobufs instead of JSON.
