I am greatly simplifying and reducing the complexity of everything happening in the world that can and will affect Bitcoin to get to the meat of calculating the current market rate for processing transactions. Luckily, the economics at this layer can be reduced to simple supply and demand—even if chaotic elements abound in the greater ecosystem. So while things like Ordinals can add demand, it still reduces the answer to this: the cost of a bitcoin transaction is whatever the market is able and willing to pay for the limited and prioritized space in blocks at any given time.

It’s entirely possible that demand will stay high for a long time, and rates could maintain 50+sat/vB or higher, but there’s a natural redistribution of those fees into consolidated miner payout UTXOs that then get sold into the market and then a reprocessing of those UTXOs by miners to mine blocks (paid for by the market). If the market is unable or unwilling to pay, mempool space clears and fees drop. This has happened several times in the history of mining already.

This does bring up an interesting attack vector though regarding censorship. If a powerhouse like the NSA moved quickly enough before lightning network really gains popularity, they could eat the cost (printed money) to inflate the mempool with high fee transactions in order to choke out smaller paying transactions (weaker currency holders in developing nations). However, this would be quite costly to maintain over time and ultimately I think the market would mitigate high fees with the kind of consolidation market behavior I described. Still fun to think about how a massive nation with imaginary money could try to attack it :)