Yeah, Mempool backlog currently sitting at 80,000 Transactions and holding steady:
https://jochen-hoenicke.de/queue/24h.html
Even if we do make new all-time highs, or whatever the price is, it would be better without the transaction bottleneck.
We can think of every transaction as a transfer from someone who subjectively values the coin less to someone who values the coin more. As coins shift to those who value them more, the total value increases. (It is also possible for total value to increase without coins changing hands if current holders change their own subjective valuation of coins. But this has more limited potential for large gains). To achieve large gains in valuation, Bitcoin holdings must move from the small number of current holders, to a much larger number of future holders who value each unit more (and each hold proportionally fewer units). Transaction constraints hobble this process.
One can also think of coin distribution as a special case of Coase Theorem. We are starting from a non-optimal initial distribution and moving towards a pareto-optimal distribution. (of course never getting there, but heading in the right direction). But for this to work properly requires low transaction costs.
We can think of the transaction fee as a floor on the possible value gain from each transaction. For example, with a transaction fee of 50 cents, a transaction that would increase the total subjective value of both holders by less than 50 cents (and thus increase total market cap of Bitcoin by up to 50 cents) will be priced out.