Negative power is still power. A negative 1,000 volts would still shock a person badly. The problem lies in simple Dv/Dt equations where you sum all values. For a sine, square or triangle waveform the long term result of say 1,000 cycles that are symmetrical around zero volts, the sum is zero volts or a fractional value < one.
For real power, not just numerical polarity, you must use the {ABS} function to include negative values as having a positive potential. Real power. In your case 6.5 uS of it.
Now your average power is the sum of the RMS values during period Dt. In cases where voltage and current are not in phase you still use ABS value and sum them together, correcting for phase angle or offset.
If this were put through a ideal bridge rectifier it would give you a single instantaneous ABS peak potential for a given instant in time. Add a capacitor to integrate over time and you have a crude average of the power, both voltage and current combined.
Notice I did not elaborate over instantaneous power. Many engineers consider it a marketing term, like for car stereos. It is an arbitrary term where you can imply high power levels over a very short period of time, but this power level cannot be sustained without damage and is far below allowed average sustained power, and much less joules.
NOTE: Whether you need them or not the other members felt my answer was incomplete, based on the below comments. So I am adding a few simple equations here that were in comments. Q & A's are archived over time, but not typical comments.
A) ( by mkeith) Instantaneous power is V(t) * I(t). RMS power is root of the mean of the square of instantaneous power. It is mathematically reasonably well defined.
B) ( by immibis) IIRC rms(abs(V)abs(I)) gives "apparent power" - "real power" means abs(rms(VI)).
