Thanks for the reference. ‘Emergence of scaling in random networks’ (1999) seems like a good starting point, and the author is sharing a copy of the version printed in Science on his website: https://barabasi.com/f/67.pdf
Is it Albert-László Barabási's conclusion that centralization is inevitable?
There are different kinds of inevitable. In systems with preferential attachmnent dynamics, hubs grow. You can break them up, but growth (and centralization) is the natural tendency barring any other force. Barabasi doesn't form conclusions on this, he shows the math and articulates the process.
I don't know what different kinds of "inevitable" exist, but I take it to mean unavoidable; inescapable. In that sense, that there is a natural tendency barring any other force towards centralization doesn't mean that it's inevitable.