So to pin your hopes on sidechains and keeping the 1MB limit is just silly. Honestly I would love to see a hard fork to add SPV proofs and fix some low level plumbing of the main chain at the same time but it may never happen. It certainly isn't going to happen in a scenario where users are finding it increasingly more difficult and expensive to use bitcoin for their transactions.
I find expressions of certainty regarding the future as cause for suspicion, but I think you sincerely meant this casually, and are thinking it is "unlikely" rather than "certainly isn't going to happen".
Certainty of the future is typically a 'tell' for an unsubstantiated argument by elevating it apriori, and ought be used cautiously.
Consider whether others may reasonably see an opposite future? One in which the more expensive transactions, especially those where the higher data size is paid by higher transaction fee, may favor high value transactions (such as an SPV proof may be) over a lower value transaction.
The loser in the block size scarcity is not going to be the SPV proofs so much as the microtransactions because the fee may swamp the value of the transaction. Things like pay per click adverts and changetip cashouts would have to raise the minimum transfer.
As much as I would like Bitcoin to be more useful, it is not worth introducing new unchecked risks.
The block size expansion does increase the trust placed in miners, there is not any getting around this. Of all the places to put trust, this is likely the best, as it already exists. Even so it MUST be done with extraordinary care. Increasing necessary trust arrangements with Bitcoin is not to be done at all if possible, and only if necessary then done as minimally as possible.