If you take hobby remote control cars as an example, I just bought an average lithium powered one that can do 45A at 11.7v, or roughly 500w to drive a <5lb toy car (very quickly). A 12lb skateboard and a 180lb rider will need several thousand watts peak I would imagine, at least.
If you're sprinting or accelerating hard you could easily be putting down 800-1000 watts or more.
Electric motors have a far greater range of RPM at which they can deliver a sufficient power.
A motor that delivers 700W attached to any gearbox will deliver 700W after the gearbox (minus losses). The difference is that shaft will rotate at different speeds and with a different number of Newton / Metres. A slow shaft attached to a 700W motor (at full power) will have more torque than a fast one.
If that's what you're saying, you're making several false assumptions and your statement is incorrect (because electric motors DO have a medium high RPM range where they are much more efficient than when near stall). Although it is true that (for some but not all types) of motors the stall torque is very high, that uptick in torque does not and cannot rise high/fast enough to balance the loss of work output caused by choosing a gearing that results in a slow motor shaft RPM. Recall that work is force over distance. A completely stalled motor is doing no work yet consuming prodigious amounts of power (which all goes to waste heat). In order to produce "700W of mechanical power into the gear train", a motor in a stall or very near stall would have to produce infinite torque, which is clearly impossible. Furthermore even if you consider, not a stalled motor shaft, but one turning below optimal RPM, you don't always get twice the torque when you halve the speed, you get maybe a fraction of that. So there's still a power/efficiency curve. Finally, a motor that is turning at an RPM significantly slower than its optimal rate is usually drawing much more power than it will for the same input voltage if running at higher RPM, so looking just at motor RPM vs torque at the shaft as a measure of output power ignores the important fact that you're putting a lot more electrical energy into the motor to get that output.
The only way in which your comment makes any sense would be to talk only about torque and say that for most types of motors, if the motor has enough torque to turn the drivetrain at operating RPM then it will almost certain have enough torque to start the drivetrain from a dead stop. So the motor doesn't necessarily need changeable gear ratios in order to get going if you ignore or amortize the energy wasted getting up to most efficient operating speed.