All variable frequency drives using power electronics are essentially "frequency choppers". Either the VFD is a current-sourced inverter topology, or a voltage-sourced inverter topology. To obtain the desired output (usually defined as voltage at some specific frequency), the incoming signal gets "chopped" and rectified to produce a DC pulse, which is then inverted back produce a "sample" of a fraction of a sinusoidal waveform.
The sample can vary in both amplitude (voltage) and duration (frequency) - which means it really isn't going to look like a sinusoid at all. However, when a lengthy series of these samples get consecutively strung together, the resulting waveform CAN look suspiciously like the traditional sine wave.
The essence is shown in the above conversation - basically you have a digital device (on-off) replicating an analogue (gradually changing continuous signal). The higher specification of the digital device the faster it can switch and deal with higher potential difference/currents.
Many VFDs output are pulse width modulated (PWM) so that over a cycle it is close to a 50 Hz (or 60 Hz) sine wave. If you look at the output with a fast storage scope you will see distinct on/off blocks.
The reality is they are close to a sine wave, but not identical. So at low specification the resultant AC wave has a base frequency of 50 Hz but a lot of harmonics and distortions.
NB - if you do a full analysis you will see not only a lot of different frequencies, but also some interesting voltages. Hence a lot of the discussions about damage to insulation, some loads won't run on "square waves" and cause circulating currents (i.e. in bearings etc.)