There does seem to be an assumption that designing a 6th Gen top end drone is somehow different to designing a 6th Gen aircraft.
At the top end they’re likely to be manned/unmanned airframes with similar range, engines, sensors etc. Yes, you can take out the seat and the man to reduce size, increase range etc - but if you reinforce the structure for higher G you add more back. Look at the NGAD where they are looking at $500M per manned aircraft controlling drone wingmen - which will probably resemble Tempest in most regards.
At the lower end you can design with one use engines for kamikaze drones such as in Ukraine, but that’s not really a viable export market because everyone will be doing it.
So the dividing line - and cost savings - is far more nuanced than being suggested.