So Bitcoin Cash is not supporting SegWit, it boasts an 8mb block size (but is that really a good thing?). This allows the blockchain to handle more traffic, however you would need about 250mb block sizes to really scale up to a global level (7 billion people doing avg 20 transactions /day) - so another hardfork would be needed, and another, and another. Only super-scale mining giants could service that size of blocks - aka, centralised.

I also believe 250 Mb block sizes would be very bad. Seriously, we need second-layer scaling options. People have become so fanatic in this debate that they aren't able to see the two sides of the debate anymore - there are for sure quite some people on the "bigblock" camp that seriously thinks online scaling is the only right thing and that gigabyte-sized blocks aren't problematic. We do need second-layer scaling, and we do need bigger blocks. Now, SegWit is not the only way to achieve second-layer scaling. The alternative proposal is FlexTrans. From my perspective, it seems to be a much cleaner approach than SegWit; while SegWit is a overly complex workaround, FlexTrans seems to be a clean and simple fix.

I see no valid arguments against an 8 Mb block size. I could have bought hardware for running several full bitcoin nodes for the total amount of bitcoin fees I've paid due to the current capacity strain, the "8MB blocks are bad because we cannot afford the hardware for running nodes" is bollocks.

Before yelling "centralization!" or "decentralization!", try to define the terms. Yes, bigger blocks do pose a centralization risk, but so does high fees! Forcing everyone to use lightning is also very much a centralization risk, it is an extremely likely outcome that almost everyone will be pushed into connecting to a big payment hub, and it's likely that there will only be a handful - maybe even only one or two such big payment hubs.