Isn't it more than that though? It isn't just about sample rate and resolution which for a single sound should be extremely accurate, it's about replicating the effect on the sound as the knobs are twisted and sliders moved around. That's a whole different set of problems and I imagine very very difficult to accurately replicate/synthesize.
Here's an example: https://www.u-he.com/cms/diva -- it's widely regarded as one of the best sounding analogue emulations (mostly due to its filters) and as a CPU hog (in its "divine" sound engine mode).
Mind you, in a modern CPU you can still run a few instances, but very fewer compared to regular plugins.
