The more I think about it, I'm not so sure there is any risk to security. Some basic points (and I am not recommending anything here, just observing and hypothesizing as a way of trying to understand what the real issues are, if any):

Mining is a competitive business. This means there will always be pressure to stay profitable. This pushes miners to cheaper energy sources and other good stuff (as often cited by Bitcoiners) but it also means that miners who can't compete will constantly be going out of business, just like in any other sector. This is inevitable, no matter how much money is thrown at miners through fees or block rewards. The question is, Is it bad for Bitcoin if unprofitable miners go out of business?

"Hashrate go up" seems to be an oft-cited expression of Bitcoin's security requirement, but is it really so? The difficulty adjustment works in both directions. When China miners dropped offline, was it bad for security?

Centralization offers efficiencies, and therefore businesses of all kinds tend to optimize around centralization. There is not much stopping the mining industry from centralizing. There is concern that heavyweight mining pools are becoming too big, which translates to having too much say in picking and choosing which transactions go into a new block. Hence, Stratum V2. However, allowing miners (instead of pools) to pick transactions only means that the centralization lines will be drawn around miners rather than mining pools, which does not stop centralization per se. The question is, Is centralization of mining bad for Bitcoin? This is where my knowledge falls short, but it seems to me that as long as full nodes are validating the mined blocks based on their own consensus rules (which are NOT centralized), then why care if mining is centralized? Won't nodes reject newly mined blocks if they don't meet the requirements of the network?

In conclusion, it seems to my thinking at the moment that the root of any theoretical security issue is not in mining fees or centralization, but in the ability of full nodes to reject blocks that thwart Bitcoin's neutrality. That is to say, new blocks should only be accepted if they follow the free-market principle of including the most valuable transactions. A block that ignores market-value transactions should not be accepted. Again, my knowledge falls short, but if Bitcoin full nodes already do this, then where is the supposed security issue with centralization? For all the network cares, a single govt could take over the vast majority of mining as a service to the network without any problem, because the miners have to do what the "people" say with their validating full nodes, right? (Then at least the govt would be serving the people as it should.)

(Again, I'm not recommending anything here, just observing and hypothesizing as a way of trying to understand what the real issues are, if any.)