Help me understand. The resources used in PoW is a function of the income acquired from mining. If we were to assume that the price remains stable, then why wouldn't a capacity increase result in an increase in the fees. The main deterrence with Bitcoin achieving higher utility is only with the capacity as it is closely related to the costs per transaction. The transaction in Bitcoin is highly elastic, a change in the fees would result a greater than proportionate change in demand. If you were to accommodate more TXes per block, then it is not out of the realm of possibility that you will one day be able to substitute the block rewards with the fees.

The computational power is not a good metric to compare with the security of the network, you will need to be looking at the costs of it. It is also not a good metric to compare using the previous halving. I'm estimating the block rewards to at least, currently be comprising of 3/4 of the miner's income at best, perhaps 4/5 at worst. You will see a smaller decrease in the drop in hashrate in the future, because fees start to compensate for the decreasing block rewards.

It has nothing to do with how Bitcoin's pricing being volatile or on an upward trend. There will be a point, in the future where the hashrate will stabilize, or at least have negligible change due to it. Fees are supposed to balance out the block rewards in the long run, you will absolutely be able to mitigate any issue with regards to the security if you are able to and want to increase on-chain capacity.

Even with the fees being a very minuscule part of miner's income, with the prices and improvement in efficiency of ASICs factored in, I'm still seeing a relatively stable hashrate throughout the past few halvings, latest halving being 2.74x. Again, think about how much better we can cushion it if fees were to form a larger percentage of the income.