You’re massively underestimating what it takes to get from an esp32 hobbyist drone to a weaponized drone with 10 mile range and an actual explosive payload capable of taking out armor (in any number). Or the sensor package it would take to make them useful against personnel. Let alone deploying ten million of them in a real war.

And you’re entirely ignoring the very real problem of the mothership which has to survive to get within ten miles of the battlefield, unless you’re planning on releasing them from box trucks which means their range will either be useless or they’ll get taken out by bigger, more expensive loitering drones the second they’re spotted. War is antagonistic co-evolution in its purest form, these naive solutions dont last very long which is why our weapons cost so much (for everyone, not just the west).

When you spread the risk across 10M units you are better off compared to placing all of your bets in one super fancy unit. Remember in the Ukrainian war, Russia took out most of the Ukrainian planes in their hangars, before even they took off.

I totally agree with you that drone swarms is not a silver bullet, and likely some effective adversarial strategy will be developed(jamming, attacking motherships etc), but the point is that airplanes are not as important as they were in the past. Ukraine is still standing with no real air presence.

The battlefields in Ukraine is how we know the real cost of those drones - they managed to push it down to around $300/unit for FPV ones, which are the primary armor killers.