There is no such thing as "speed limitation". Automatition technologies seems to not know what they talking about in this case.
Also they never said to us that the had any problems like that, so it is very strange that they saying that...sounds like they have higher profit on selling smoothsteppers which might justify such opition.
Mach3 has issues with high feedrate operation though, because of it's errorous trajectory planner, but there is no motion controller which can fix that, because the issue is in the trajectory planner algorithm of Mach3 which can't be fixed by a motion controller plugin. The UCCNC has no such issue.
The feedrate is limited by the kernel frequency and the steps per value of the axis, those define the maximum feedrate, because the kernel frequency is the maximum rate of step signals the controller can output per axis and the steps per unit value sets how many steps the axis needs to move one unit distance.
The kernel frequency for the UC300ETH is 400 000 Hz (400kHz) which means that the controller can output 400 000 steps per second for every axis.
That should be fine for any stepper drive, because optocouplers in stepper motor drives also have limitations which is usually lower than 200 000 Hz as high speed optocouplers are expensive, so drive manufacturers not bulding $2 optocouplers into their devices, but instead they built $0.2-$0-5 ones which can do much less frequencies than what the UC300ETH can do.
Servo drives can often handle higher than 400kHz step freq., but usually that is not required because the encoder resolution and the max. rotational speed of the motor is not that high.
The smoothstepper can do higher frequency, upto 4MHz, but higher kernel frequency involves a risk of the signals suffering noise issues.
With 50% duty cycle step signal at 4MHz the pulse width is only 125nanoseconds. You can imagine how short those pulses are as they are in the nanoseconds range.
We don't feel that safe for a CNC machine working in an electrically noisy environment and so we have no plans to make a controller with that high kernel frequency.
If we would go to that high frequency we would use different protocol than step/dir, e.g. EtherCAT, but that would then require a drive with EtherCAT interface and ofcourse some more development from us to make it work.