Yes, Segwit will effectively allow larger blocks, but that's not its main purpose, and the increase is marginal.
You argued for a 50% increase yearly. Segwit represents a 400% increase.
I want to be reasonable. If we could find a way to - I don't know how - reduce the size of transactions to fit more inside a block, I'd be happy, but if we want BTC's use to increase 10-fold, then 100-fold, we will have to increase the block size, and I don't see it as a big deal.
Those two statements are more or less contradicting one another, what if we could reduce the size of transactions by a factor of 10 or 100? Changing the signature scheme from ECDSA to the proposed Schnorr scheme wouldn't quite achieve that, but it's still an improvement. There could easily be other ways to reduce the size of each transaction also.
The simple fact is that on-chain scaling isn't possible for 7 billion humans, I forget the exact estimate, but we're looking at each block growing to 100 GB+ to accommodate that kind of growth. Even 1 GB blocksize is totally impractical, 1GB every 10 minutes isn't feasible for anything but the fastest fibre broadband. 100GB every ten minutes isn't possible for any domestic connection today.