In the absence of other evidence, its probably going to boil down to "design tradeoffs", the fundamental reality of engineering and especially that of intensely closed systems like flying machines.
If a J-20 -esque main IWB is sufficient for 4-6 PL-15/16, and you'd need to make a significantly larger aircraft to fit the next class of A2A missiles (see J-36 for PL-17), then there's no point in making the J-XDS IWB any bigger than a J-20 IWB. If anything, in the interest of optimising space usage, the J-XDS IWB might even be a bit smaller so long as it can still carry the same payload. The J-35/A IWB is a tad bit smaller than the J-20 after all (at least visually from the outside), and its expected the carry the same amount of PL-15/16.
And why wouldn't you make your carrier-based 6th-gen any bigger? Because carrier space and flight deck length put significant constraints on design, especially aircraft size. Fielding an aircraft with sufficient size for internal carry of PL-17 -class missiles probably isn't possible with current or foreseeable future carrier designs. Type-004 may be marginally bigger than the Ford-class, but that's not going to be enough for J-36. And assuming a hypothetical hypercarrier design does emerge before the end of the J-XDS' lifecycle (sometime in the next 50 years or so, a very long time!), at that point the original land-based J-36 should be a mature platform and explorations of a naval variant could be possible, so there's no permanent opportunity cost there.
There's also the possibility that SAC, or rather the Chinese MIC, is hedging its bets. Note that two concurrent complimentary next-gen fighter programmes (intended to be seen into service induction) is unprecedented in human history. The J-XDS already hedges against the J-36's "all-in" on maximising stealth, speed and range at the expense of WVR anything, in a similar way to how the YF-22 hedged against the YF-23. The J-36, in turn, is likely a hedge against the J-XDS which would naturally be more reliant on the emerging CCA concept to carry external payloads in the absence of a larger singular platform. So we can partially attribute the J-XDS' non-increase in IWB capacity to its experimental reliance on CCAs, of which there exists multiple possible strategies in future naval warfare (including, but not limited to, support carriers like Type-076 launching CCAs in support of manned platforms from full carriers).