I have still been wondering about this a bit, and yeah, tweaking intermediates seems feels right & aligns with how base Factorio mkes things expensive. With that in mind, a couple of other reflections (mostly) regarding the K2-specific resources. Many already seem fine, as in, appropriately more costly.
*Glass & silicon: don't need to be individually more expensive, the electronic components entail a cost jump.
*Rocket fuel, atmospheric processing, other fluid/gas chemistry: there's a lot of new gasses but I don't think any costs need to change. Base Factorio doesn't really make petrochem recipes more expensive, just makes the downsteam items more costly. Eg blue chips taking more acid, which the expensive universe also does. More acid / chlorine is already needed for the extra uranium / rare ores, I don't think anything in the K2 specific chemistry needs more adjustment.
*Powergrid / fuels: fuel recipes don't seem more expensive in vanilla, the player already has to build more grid to power more of everything else. So processes into eg tritium or fuel don't need to change in the expensive universe either.
*Wood & biomass: These currently don't change in expensive universe vs normal K2, and base Factorio doesn't really have an analogy. Maybe these buildings could just make less? Maybe half the output number, or else double the cycle length? Not a huge thing.
*Uranium: Usually this is a power source but in K2 is a science ingredient. Expensive universe centrifuges eat more uranium ore than normal, that's good way to do it. It's fine.
*Lithium: The lithium-sulfur batteries get more expensive, affecting advanced and singularity tech cards, but not matter tech cards. It's good, I don't think everything needs to get worse everywhere.
*Immersite: The expensive universe is more expensive, but I'm not sure if just needing more quarries is as interesting as it could be. It doesn't subtly change ratios like expensive green or red chips do. What if crushing raw immersite gave 4 stone instead of 3? Not a big deal.
*Matter: It's either for antimatter powergrid, which doesn't need to change because powergrid, or else whatever you're converting into will encounter expensiveness later. I wanted to mention I considered it, but it does already seem fine.
So, having tried to be comprehensive, I'm concluding that pretty much everything is already good. Maybe a note for wood or biomass, but they're not a big deal.