Seems to me that Adobe is building in functionality that is processor/gpu intensive.
This is a trend across the software industry. Other photographic software is doing the same, Adobe is simply keeping in step with what's going on generally. Part of the point is to make more things faster, which is what we all want. CPU processors , especially from Intel, have stagnated in recent years, so instead CPU makers added cores. GPU power has advanced much more quickly, but that power is not fully used by most software because it can take some work to do properly. Recent major versions of Lightroom, Photoshop, and other vendors’ applications have tried to take better advantage of both the additional cores and GPU power that are available in the computers you can buy today.
Also, the new Enhance feature is powered in part by machine learning (ML). Again, this is a general trend in computing. The potential of machine learning/neural networks/AI is so high that both Windows and macOS have implemented machine learning support at the OS level in recent system updates. If you use the new Neural Filters in Photoshop, those OS versions are required. The new Apple Silicon in Macs includes a hardware Neural Engine for the same reason: The entire industry expects applications to increasingly use ML/AI/neural capabilities from this point on, because it is useful in so many ways, and the hardware support is now there to make it happen.
I just hope they do not go too far and we end up with a bloaty piece of software containing functions that many if us hardly, if ever, use.
It’s hard to think of why Super Resolution would be bloat. It has attracted a lot of interest, because many photographers are interested in making larger prints from tighter crops or smaller sensors. Like a lot of ML-based features, the time needed to do the required calculations is impractical with older computers, but has now become practical with the multicore CPU/GPU/ML power we have today. Before Super Resolution, many Lightroom and Photoshop users paid an extra $99 for Topaz Gigapixel AI to do the same thing. Now, there may be a fair number of potential Gigapixel AI customers who will decide that the one in Camera Raw is good enough, and choose to save that $99.
And the performance improvements from optimizing for current CPU and GPU power certainly have universal appeal, something everyone wants to use.