Also isn't monero scability worse than BTC (ie. more data per transaction to process)? (I know Monero has an adaptative blocksize but that doesn't solve the centralization of nodes problem).
The difference is modest, now that people have adapted to using Monero efficiently (not sending payments with lots of unnecessary decimal places for example). Go look right now. Many real transactions using ring signatures (now required) on the live network are <1 KB. Some are larger of course. Median block size last 800 blocks 3500 bytes, average transactions per block 3.5. Average tx size after subtracting block headers is <1 KB.
The bigger increase in transactions size will come with the (pruneable) range proofs of RingCT. The same issue would affect BTC if it added CT and if not then achieving any kind of privacy (though not as good) requires engaging in mixing, chained coinjoins, etc. which reintroduce the size penalty in the form of more transactions.
There is no free lunch.