Oh, bloody hell, here we go again....
OK. Typical MOSFET designs are current output devices - that is, they look, to the sub, like a Voltage Controlled Current Source. The controlling voltage is the input signal.
We can model this, roughly, as I(out) = g*V(in), where g is the gain of the amplifier. Note that for 99% of amplifiers, that's actually a fixed value, and what we consider the gain control is actually an attenuator on the input signal.
OK, let's take a hypothetical amplifier with g = 5. We apply a 1V RMS signal on the input. Since g = 5, the amplifier attempts to force 5A through the speaker.
If the speaker is, say, a 4 ohm sub, this results in a voltage of 20V across the output terminals, for a total output power of 100W.
OK, what if we swap in a 2 ohm sub? Now we see a voltage of 10V across the output, for a total output power of 50W.
So why can an amp deliver more power to a 2 ohm sub than a 4 ohm one? Well, that's down to the power supply. If the switchmode PSU in the amp runs rails of, say, +/- 50V, it can swing an output signal of (nearly) 100V peak to peak. That's an RMS voltage of about 35V. Since this is the largest voltage it's possible for this amp to swing, it has a peak theoretical output power of only about 300W into 4 ohms, but 600W into 2 ohms.
Upshot is, if your amp is rated to 1000 watts into 4 ohms, but with the "gain" at full you aren't swinging 178V peak to peak (63V RMS) across the outputs, your amp still has more left in it....you just need to crank up the signal source more.