I think SQL/PGQ have the potential to outright kill the whole field of PG databases (if/once existing databases actually start implementing it).
The big marketing claim of Neo4J & friends has always been that they make graph queries possible, with a big portion of that claim being that those queries would be intractable for RDBMs performance-wise.
With SQL/PGQ it becomes quite apparent that there is next to no magic sauce in PG databases and that it's all just syntactic sugar on top of node/vertex tables. All the rest of their query plans looks 1:1 what you'll find in a RDBMs. Now that they are on a level playing field query syntax-wise, PG databases will have to compete against RDBMs with a lot more advanced query optimizers, better administrative tooling etc..