Or probably the readout is very accurate and we are not used to accurate readouts ...
Most readouts just measure voltage peaks and display an "equivalent" power computed considering a resistive load of 8 ohms. However some modern designs sample permanently the voltage and current and display the computed real power of peaks - considering that many current speakers have some points of low impedance, they show values much higher than can be expected. When I hosted the DartZeel NH458 I also found that the readings were sometimes much higher than I could expect from my system calibration - the SoundLab's have very low impedance in the high treble.
But how much energy is in the high treble? Very little I would say. Perhaps it had more to do with the reactive nature of the speakers? Afterall, nearly all the power that is going into a Soundlab is coming back into the output of the amplifier (back EMF) because the speaker behaves more like a capacitor rather than a resistor. Not much is lost as heat, unlike a normal voice coil speaker, which heats up and this is energy lost to heat. That back EMF could possibly be read as power by the monitoring circuit as it circulates through the local feedback loop in the output stage (I think darTZeel only uses local feedback at each stage and no global feedback). For sure it imapcts the usable power of the amplifier.
Also, I think the voltage sensitivity of the Soundlabs is quite low, is it not? I would expect a very different result with your Wilsons for both of those reasons. BTW., do you know what is the highest impedance of your Soundlabs? I had a pair of STAX ELS-F81 that were like 60 ohms in the bass up to 500 ohms in the mids, down to below 4 ohms in the treble. They were a nightmare for most SS amps...but a big 100 watt Pushpull Class A amp sure sounded good. My Acoustats though, were pretty moderate except the highes treble (2 ohms) and worked well with all types of amps (except one OTL that went into oscillation).
