Have you ever been involved in someone teaching a lecture from a "script"? I don't know if there is a specific English term for this, I mean that the teacher/professor collates the material beforehand without publishing it in a book form, and makes it available to students.
If you see this being done year afer year, you will see what kind of changes are made.
The importance/length of sections is changed relative to each other, due to changing emphasis, the need to make space for new material, and the time needed by the professor and/or students to get through a section
practical problems are changed to be more understandable, new problems are added, etc.
material that was difficult to bring across is rewritten to be presented in a new way
small new discoveries are mentioned, such as using the results of nifty new studies as examples that emphasize a point
corrections are made, since there are usually errors at the beginning
I would say that textbooks go through similar changes between editions. This is entirely normal - a large and complicated artefact like a textbook is best created in an iterative manner, not unlike a software programm.
Major discoveries in the field are much less likely to trigger a new edition. First, the future impact of many discoveries is not recognizable when they are made, and they linger in some small journal before the discipline notices them and makes something out of them.
Second, even when something is recognizably new and different, and excites scientists, it is still not "fleshed out" enough to be taught to students, since it doesn't yet have its own ecosystem supporting literature, successful application in large projects, whatever. Third, the kind of professor who gets to publish a textbook is usually old, experienced - and set in his ways. If he dedicated his life to building superconductors out of metal alloys, and some young upshot shows that graphene can be used in a superconductor, the professor will wait for a few years whether that new technology (which is in direct competition to his own research) will establish itself, before starting to give it space in his textbook.
I think there are a few exceptions to that "discoveries percolate slowly into textbooks" tendency, for example I heard somewhere that CRISPR/CAS entered general genetics textbooks rather quickly. But it is much more typical, especially in undergraduate level textbooks, that changes between editions are incremental improvements of existing material.
There are also some fields where the changes are very impactful. This happens in fields which study human-created rule systems, typically law, but also accounting. In a law textbook, a subset of laws changes every year, and their interpretation by courts also changes with new case decisions. The new editions of textbooks have to reflect these changes.
A reason for professors to want the newest edition (beside monetary ones) is simply that it makes it easier to teach. With a class where the students use multiple editions at once, there will be difference in the text, but especially also in the problems. Making sure that everybody reads the same text, or is solving the same problem when homework is given, is a huge headache if students use different editions.