If you're sure of the output power rating of 94 W, and we assume that it's measured at the customary few percent THD or onset of clipping, here's how the math would go, working backwards. To simplify the math, let's assume the output power is 100 W. Since P=E^2/R, E = sq rt ( RP) = sq rt (100 x 4) = 20 Vrms, which equals 20 x (2 x sq rt 2) = 20 x 2.828 = 56 V pk-pk or ±28 V pk-pk. If we assume the output stage clips at voltages 2 V less than the DC rails, then the DC supply would need to be ±30 VDC. This is consistent with a raw power supply using 30 VAC on either side of the power transformer center tap. Now, how much DC current is required for the output stage? Class AB or class B output stages have a theoretical maximum efficiency of 78.5%, but real-world circuits run around 65% (that other 35% is what makes the heatsinks hot!). So, to get 100 W of AC output power, the output stage needs 100 W/0.65 = 154 W of DC power. Since there's other small-signal circuitry, indicator lights, etc. that may consume 6 W, total DC load on the power supply would likely be 160 W at full output. With total rail voltage at 60 VDC, and since P = E x I, I = P/E = 160 W/60 V = 2.7 A. Allowing for a typical 90% efficiency of the rectifiers and typical DC resistances in the power transformer secondary, that would translate to roughly a 3 A rating for the power transformer secondary. Just for added information here, if we assume the power transformer has losses of 5% to 10% (what makes it get warm), then the actual total input power to it is only 5% or 10% higher than the 160 W DC output power. At 170 W input AC power, the line current at 120 VAC would only be about 1.4 A. The "rated" power of "300 VA" is related to a worst-case consumption, usually with some sort of fault conditions, for fire safety considerations. Likewise, the fuse rating is likely 4 A because UL considered it adequate to prevent a fire under worst-case fault conditions ... and Fender likely wanted the rating high to prevent turn-on inrush currents from popping the fuse. Usually "slo-blo" time-lag fuses are specified for that purpose, but Fender likely also knows that, in a pinch, people will install any type of 4 A fuse, including the faster blowing standard type ... so 4 A will save a lot of annoying phone calls. But neither the fuse or VA rating gives an accurate picture of actual circuit operation. I hope this helps!