Rampant speculation:
All of this is I think the reason why we will inevitably be fielding autonomous or semi-autonomous automated platforms for our cutting-edge, high-tech warfighting against similar tech opponents. The cost of maintaining a small, bleeding edge tech level force for worst-case scenarios, in addition to a much larger lower tech force for conflicts that are much more likely to happen is getting prohibitively high. The best platforms are overkill for the most likely scenarios, and the potential risk of losing even one of them limits potential use cases dramatically. Yet the costs to not just develop but maintain and operate them to preserve their operational capability has to be paid even if they are never used.
Now, replace most of those systems with autonomous or AI operated platforms. You don’t have to train physically, it can all be done in software except for periodic validation of mechanical functions. You don’t have to constantly hone a human’s skills and reactions. You can store the systems much more efficiently and cheaply without affecting readiness. Unit cost may be high, but ultimately should be much lower than human operated systems, as you don’t have to engineer in all of that stuff to protect the meatbag. Best of all, maybe, you can actually use these things in a wider variety of circumstances because while there is technical risk in someone getting a hold of one of the systems, there is no risk to a human operator and hence much less political risk.
I actually think it is more likely we will push automation at the high end than at the low-intensity conflict level. There are tons of problems with, say, using drone swarms against insurgent camps in a forest in Iran, in terms of targeting, collateral damage, and public relations. There are far fewer of these problems with, say, fielding a flock of autonomous air superiority platforms to counter an enemy air attack on a US naval task force or Taiwan or whatever.
Finally, For all of the talk about airplanes and ships and stuff, I think that the real growth area may well be in high-tech ground warfare gear, particularly infantry related equipment. Infantry is the basic common denominator for much of what we have been doing in our wars the past several decades, and the multiplicative impact of things that enhance soldier survivability, sensing, and firepower at this level should not be underestimated. You are going to need boots on the ground, and if you can field systems that make each soldier do the work of four or five or more current soldiers, that’s a huge force multiplier. This does not, and should not, mean automation necessarily. Some things could benefit from autonomous systems, like mine/IED clearance, scouting, defensive systems perhaps. The basic job of the infantry though is harder IMO to automate than that of air superiority pilots.