Its also easy to naturally batch things. When the load is low the batch size is 1, and then it naturally grows as more work gets queued during processing of the prior work.
I love ring buffers and queues.
e.g. https://blog.cloudflare.com/cloudflare-architecture-and-how-...
The majority of the cost (~90+%) is actually due to software. In the system call case this is the implementation of the system call. In the context switch case this is usually the cost of the scheduling decision. These operations take material amounts of time and access memory which evicts user caches. It is in speeding up the software component that real gains can be made. This is not to say that system calls could ever be made "free" since function calls/loads/stores will still be ~100x faster, but the cost would probably be ~10% of the current cost if they software component could magically be made instant.
As an example of what I am talking about, there is this talk by Google from 2013 [1] which describes a "fast context switch" system call (benchmark at 18:40). This "context switch" knows what it is switching to which lets it bypass the scheduling decision in normal context switches. As a result it takes ~170 ns (which is pretty close to the hardware cost mentioned in the OP) vs 2000 ns of a normal context switch.
I could go into more detail as to some of the causes of the underlying software costs, but that is a discussion for another time, so I will just leave this off with summarizing my point.
tl;dr: System calls are slow. It is due to the software implementing them, not the hardware.