That is a good point; a back of the envelope calculation for aircraft prices (aircraftcompare.com) means that for 5 billion USD nowadays you get about:
- 36 F-22
- 25 JSFs (@200mil. piece, debatable)
- 74 F/A-18E
- 71 Typhoons
- 73 Rafales
- 81 Saab Gripen
- 106 F16s
- 138 strike eagles
- 250 A-10s
- 185 Mig-29s
- 156 Su-27UBKs
- 76 Su-35s
This is a very crude estimate (no fuel or maintenance costs, which would make these figures even more extreme.)
Still I believe if you have 156 Su-27s or 73 Rafales operating from stretches of straight roads disseminated in the countryside you could wage a much more efficient air war than if you have 36 F-22s. The F-22s probably would have a hard time ensuring a 24-hour air patrol (except if you are covering the area of switzerland...).
Plus if you take a head-to-head engagement between 30 F-22 and 150 sukhois, a number ratio of 5, the F-22s need a 62% missile success rate to shoot all sukhois down, but then would have no missiles left for bombers or whatever the sukhois were escorting, thus the risk of losing F-22s on the ground.
In general, I get the feeling that we are in an analogous situation as just before vietnam, when military thinkers though the age of dogfights over and did not see the use for a cannon in a fighter. See how that turned out. Why should it bee so different now?
As for totally autonomous drones, I do not think it is possible today to have a failsafe and efficient way of distinguishing friend and foe, meaning that some kind of targeting datalink will be needed.