Size of blocks needed for scaling?

Unless BCH is planning on adding some layer 2 technology what amount of population could different blocksizes support.

If on average 1mb blocks support 3tps it gives 260.000 tps per day.

If everyone used bitcoin for all their daily payments, and on average people made two payments per day, lunch plus something else, that would make it possible for 130.000 users with 1mb blocks if all of them were stored on chain.

If however Bitcoin is used as a store of value and only used when needed and for large payments, and people would only make large payments on average twice a month such as paying rent, or buying a car a new phone or making transaction from an exchange etc.

Bitcoin with one 1mb blocks could at most support 3.9 million users if all of them were saved on chain.

So the amount of users a blocksize support is depending on how Bitcoin is used.

With 32 mb blocks the numbers of the two different use cases would be:

4 million for two daily transactions on chain or 124 million for two monthly transactions on chain.

With 256 mb blocks.

33 million users with two daily transactions on chain or 998 million for two monthly transactions on chain.

You would need 2.5 gb blocks in order for Bitcoin to be used for two daily transactions by 330 million users.

How is using Bitcoin for two daily payments an average and storing them on chain sustainable long term if the block size for more than 330 million users have to be more than 2.5gb blocks every 10 minutes?

I still think 1mb blocks are too small and think blocks could be up to 256 mb today and for a couple of years. But I don't see 2.5 gb block size as sustainable before the cost of storage decrease by a factor of 10. And 25gb – which could still not support everyone on earth if every transaction were stored on chain – before both storage space and read write speeds increased, decrease by a factor of 100 and that might take decades?

How are you thinking this would be solved?