And both are the reason why German universities rank so badly in international rankings. They are attached to the universities, forming clusters, but their accomplishments are not measured anywhere.
This suggests that: a) German universities don't care that much about international rankings, and b) organizations doing the international rankings don't care about their accuracy - otherwise, they would have adjusted their rankings to compensate for the idiosyncrasies of German universities.
More precisely on (a), the unis care about rankings, but the politicians who'd have the power to change the structure of research don't. Indeed for the majority of German politicans, the word "elite" is toxic, even and especially in the context of having a few world-class centres of excellence rather than 50 merely adequate universities.
Global Uni rankings have a big bias toward research. What the poster is saying is that Germans do more of their research at these institutes than at their Universities so German Universities rate poorly on global lists.
As you can see, in Germany and France, research institutes take the place of what in the US and the UK would be universities like Harvard, MIT, Oxford, and Cambridge. Fraunhofer doesn't even show up, because while they do a lot of amazing stuff, they aren't under publish-or-perish constraints (they instead have contracts to fulfill) and so they publish comparatively little.
This has a number of reasons, some of which are pragmatic. For example, if a French or German university wants to hire a professor, that professor usually has to be able to speak French or German, respectively, at least at a near-native level to be an effective teacher (there are exceptions, but it's still a frequent requirement). Research institutes aren't constrained in this fashion, even though they closely interoperate with universities.
For pure research, Germany has its Max-Planck-Institutes.