I remember early on during 6th gen discussions there was a lot of talk about it being optionally manned. I think some specialized unmanned variants of the J-36 could fulfill some valuable support roles. It would probably be a waste to just make them into unmanned weapons platforms like what a CCA is typically imagined as, you can probably do that with cheaper, specially designed platforms. But as unmanned force multipliers I think its high cost can possibly be justified. For example, you can make the cockpit and the IWBs into swappable modules. Place fuel modules in them and you can enable very deep (e.g. 5000km+) or fast (e.g. supercruising the entire flight) missions. Place EW modules in them and mass a few of them per each manned J-36 and you can have a distributed EW spearhead ahead of a large strike package. They would have essentially identical flight profiles as the J-36, thus greatly simplifying operational planning and logistics while increasing flexibility. Of course, you can also just place weapons in the IWB.
Taking a more generalized view, it would essentially make it so that all J-36 missions are solo missions as whenever you need more than one you can just add more unmanned versions to the formation.
‘Optionally manned’ design suggestions weren’t really about having hot swappable modules for pilots and other needs, it’s more like the equivalent of an F35U unmanned version built from the ground up as a CCA but sharing as much airframe and parts commonality as possible with the manned versions for lower R&D and maintenance costs while also enjoying greater economies of scale across the board.
The problem with that suggestion for the likes of the J36 and even J20 is that the cost-benefit balance tips massively against this suggestion the bigger the base platform, since the weight savings as a percentage of total aircraft weight saving/range extension drops the bigger the base aircraft weight, which in turn eats up the range increase.
Especially in the case of the J36, with it needing 2 pilots from the get-go, you will loose a massive amount of combat potential removing the pilots and just end up with a massively expensive CCA that doesn’t bring anything remotely justifying it’s cost to the table.
For China, there is also the added consideration of production bottleneck, as it will want next gen assets in frontline service ASAP, so having one factory building and common set of core components for both the manned and unmanned elements is going to slow both down in terms of peak production rates, so it’s far better to have different factories and component suppliers so production can run in parallel without as many common bottlenecks.