Electric motors generally make peak efficiency at rated RPM, so yes but there are very few systems that actually have variable motor speeds.
With large chillers specifically, it works like this:
First, if you’re comparing apples to apples, fix the condenser temp and evap temp and total btu/hr (otherwise we’re not comparing the same load/system). That means the same amount of refrigerant is being compressed across the same pressure differential. Take 1 chiller rated at N tons, and another rated at 2N tons.
The smaller unit has less friction loss, fewer starts and stops, less time at load/unload where the compressor isn’t really doing any work etc. If this is your idealized comparison, the smaller chiller is using more power to do process work and less to just “idle”
To your other points, it’s worth noting that scroll, hermetic, screw and centrifugal compressors are absolutely capable of and designed to run at 100% duty cycle. They are cooled by the process fluid and so if load is constant, they settle into a steady state temp within 2-3 minutes where they remain unless load is altered. In fact, short cycling from oversized systems is responsible for more premature failure because frequently oil can’t return to the compressor between cycles and lubrication failure at startup causes burnouts.
I love learning and educating myself, and I’m always happy to accept that I’m wrong if it means I’m better for it. When it comes to dick swinging, I normally don’t fly off the handle. I thought I wrote two very solid explanations to clarify some confusion about these systems, and got told “you don’t know what you’re talking about”. Then you retorted with information that is diametrically opposed to fact, and it set me off.
Lastly, because Huber costs a fuckload of money does not mean their engineering is spot on. We used to use them at IPG photonics in laser applications that were much more precision dependent and had less load. They still fail. Usually stupid stuff from being over featured (not a bad thing if holding 1f is worth not breaking $200k in optics), sometimes just from leaks.