Loyal wingman aircraft would have to be stealthy though to do it.
The whole point of the F-35 is that it is a stealthy platform with a very advanced set of targeting and weapons capabilities: but the issue is a similar capability drone would basically have a similar price tag since the expense of the pilot is minor in that case.
Like plausibly the best autonomous F-35 platform would just be an F-35.
Low-level drones are also stealthy. If they're not stealthy enough you can send them in bulk to overwhelm defences.
Ukraine's success with precision targeting of Russian assets would have been unthinkable with a fleet of F-35s and associated missiles because of the cost.
Cheap, smart, disposable drones are orders of magnitude cheaper and at least as effective for many mission profiles.
It's not the expense of the pilot but the fact that the human pilot is subject to human biological constraints. Ie they black out after 9 Gs or so. An autonomous platform you'd want to design to be able to routinely do outside that envelope in a way that a platform designed for a human occupant is not.
Its a nontrivial exercise to design a plane which can regularly do more Gs then that though.
There's no drones currently in development AFAIK with this goal either - the focus is on getting them to do useful mission without a pilot, rather then out dogfight or outmanoeuver missiles.
Basically the point stands: a hypothetical "better then F-35" drone is likely more, not less expensive then an F-35.
The whole point of the F-35 is that it is a stealthy platform with a very advanced set of targeting and weapons capabilities: but the issue is a similar capability drone would basically have a similar price tag since the expense of the pilot is minor in that case.
Like plausibly the best autonomous F-35 platform would just be an F-35.