TLDR; Miners will propagate their blocks before they are mined. There are several 'thin block' protocols that have existed since before BSV split. These work on the assumption that the block recipient already has most of the raw transactions on hand, and can reconstruct the complete block using transaction ids. There are even ways to only send partial transaction ids. BSV intends to take this one step further. Miners create a block template and through the process of mining attempt to solve their template with a random number. What will happen is the template gets sent to others before the miner has found a solution. Then once they find a solution, they only need to send the solution. Miners typically update the template they are working from every few seconds to include newly received transactions. Every time the template is changed, the difference between the new and old templates is sent out. This results in every miner having a copy of all the others miners blocks, before they've even been mined.

Small blockers priests will tell you how BitCoin will never work, how they proved that mathematically already (but without sharing that proof). Those people don't want BitCoin, they want what visa, mastercard, and paypal already have: closed networks that they control and can leach a percentage of every transaction from. They are rent seekers and proponents of the status quo, and they will behave badly to get what they want.

Transactions processors form a small world network. This can include direct fiber optic links between them. I'm sure some enterprising network will rise to commercial demand (Equinix?). I suggest studying how networking technology scales vs computing and storage. The network links, and hence, big blocks, are NOT going to be the bottleneck. George Gilder wrote a great book on this years ago: https://www.goodreads.com/book/show/1225914.Telecosm

Define necessary time, and who are you propagating the block to?