The problem with the third book was that it rapidly expanded in scope and grandeur but was the same number of pages as the previous books, so concepts and scope and scale which had enough exposition in the previous books simply fell short in the third.

Kind of, but I think even so, it just wasn't consistent enough. It felt like his publisher demanded, "Ok, you owe us a third book. Let's go for Stephen Baxter's scale, with Greg Egan's mathematical prowess." That's not a reasonable goal to shoot for, unless you're Baxter or Egan.