> TFA is literally about a $96 rocket

It's a firework-grade rocket with no payload that can't even ignite reliably.

To imitate even a TBM or a MBRM, you need similar kinematics, even if you're running without a payload. Maybe your solid rocket motor would be a touch smaller because you're not delivering hundreds of kilograms of explosives, but it still has to be large because of the rocket equation. With a large motor you're looking at a lot of damage if it explodes at the launch point, so you need quality casting. You can't really save much money on the motor.

Then, you need a TEL. Because the motor is large, the launcher has to be comparable to the real thing. You probably don't want to have two different vehicles, so you keep the same vehicle; it needs to be armed, driven around, and set for launch. Not that different from the real thing.

So you've done all of that, and then you realise that your empty warheads are too light and the missiles (or warheads, if you split) don't interact with the atmosphere in the same way as non-decoy missiles do. What's worse, modern radars are perfectly capable of noticing that and discriminating the decoys. All of that effort, and you didn't win anything. Might as well add the payload.

The US and the UK spent vast amount of money chasing exactly your line of reasoning with nuclear warhead decoys. Chevaline is a culmination of the effort, and it's retired for 30 years. In the end, relying on decoys doesn't really work, they are too expensive.

Fancier CPUs change very little of this calculation, because compute is a very little part of the cost to begin with.