One of my university professors helped start CSNet which ran as an alternative to ARPANet for a few years over X.25 equipment. There's a complicated history but it sorta ran until the late 80s by itself until the Internet kind of swallowed everything.

My impression from my professor is that these academic precursors to the Internet were really intended to be centrally controlled by their various committees (as academic efforts usually are). The notion that these packet switched networks would eventually just sort of be run in a mostly decentralized way didn't seem to have occurred to anybody.