Comment by mrb

A trivial solution would be to increase the block size, but unfortunately Bitcoin Core developers refuse to do it, under the pretext that it will "increase centralization". In reality, a reasonable (2-4×) increase wouldn't. The amortized cost of running a full node able to support 2-4 times larger blocks is only $5 per month (http://blog.zorinaq.com/full-node-on-5-dollars/). The block propagation delay has also largely been solved with Compact Blocks which transmit a nominal 1MB block in ~20kB. Also, at least 2 on-chain scaling technologies are being developed that would easily let us increase block sizes: UTXO commitment sets can reduce long-term blockchain storage by 50×, and Graphene reduces the size of a 1MB broadcasted block by another order of magnitude, down to ~2kB.