As a defi user / Ethereum dev, there are two things that seem relevant:
Many of the use-cases for which Kleros seems good for (e.g. arbitrating social contracts, prediction markets, etc.) is not where most of the capital is on Ethereum. Most capital is engaged in various overcollateralized lending/borrowing, LP, derivatives, synthetics, etc. So until we see additional applications which run on the social layer, I don't think the utility of Kleros will be made that apparent.
Partially related to the above, the lack of social-based applications I think has to do with a combination of generally poor UX and high latency for Ethereum. We're seeing some exciting updates on this front with rollup solutions like Optimism and zkSync which can provide ~10x increases to start, but again, the ecosystem hasn't really caught up.
More broadly, I think that attention in crypto markets shifts often, and valuations are kind of all over the place. I think I believe that scalable protocols for dispute arbitration if done well can be worth several multiples of Kleros' market cap, but also it's unclear if Kleros is that specific protocol? I personally need to dive into this more.