An informed question. I will edit the article to elaborate on this further when I have time.

The main reason is that in the current status quo, miners are rewarded with new currency — but when you shift from that to users paying transaction fees, there will be a massive pressure to keep transaction fees as low as possible and to maximise miner profits, so miners will compete against one another on costs. (Transaction fees will still be huge, because of Bitcoin’s inefficiency — but nonetheless).

There are only really a few ways in which a miner can perform the mining process for a lower cost than another miner:

1. Be in a place where electricity and cooling are cheaper (the main expenses of mining in PoW systems)
2. Have massive economies of scale (it’s a lot cheaper, on a per-computer basis, to be running a thousand servers in a warehouse than one PC in your house — partially because the electricity bill will benefit from economies of scale, but also because less electricity is being used on other tasks besides mining.)

The only miners who can meet the requirements above will be extremely rich ones, with huge amounts of hardware located in geographic regions with cheap energy. Right now that’s already what we’re seeing — to make the most money possible, cryptominers set themselves up in any country with cheap electricity.

As transaction fees slowly replace new coins as miner rewards, the cost of transactions to users will fly through the roof. As time goes on however, pressure to mine for the lowest possible cost will force that transaction fee down significantly — enough to destroy most mining operations’ profitability, but not enough to make Bitcoin even remotely good as a currency (e.g. at its current $250-per-transaction price point, just for energy — even if the pressure to provide low fees resulted in this dropping by 90%, $25 per transaction is still an absolutely *gigantic* fee and not viable for any financial system).

Right now, the only reason people don’t pay $250 per transaction on Bitcoin is because miners don’t need to charge much in transaction fees — most of their reward is from new coins. This article from 2017 states, at that time, miners would receive ~12.5BTC in mining rewards (i.e. new coins) and around ~0.5 BTC in fees; as a result, user transactions are massively subsidised at the moment, especially given that the mining process becomes harder and thus more expensive as time goes on due to Bitcoin’s design. Once this shifts to a predominantly fees-based model, due to the drying up of the supply of new coins, users will have to bear the full cost.

That means that most mining operations that aren’t gigantic and using cheap energy only available in some geographic locations would run at a loss — and nobody is going to mine for a loss when there is nothing for them to gain from it, which is why it is a tragedy of the commons. Thus, you would end up with a handful of very large miners, who could easily band together to raise prices and form a cartel, or subvert the blockchain if they so chose.

In a sense, one could argue this has happened in real-world systems (Visa and Mastercard completely dominate the card sector , with 60% and 30% of the market respectively — it’s not cost effective to be a small provider, so it’s almost inevitable that a few big players will come to dominate such a field). However, the consequences are totally different — the domination of Visa has not resulted in exorbitant prices for transactions since normal financial systems are several orders of magnitude more efficient than cryptocurrency blockchains by design.

In addition: unlike cryptominers, Visa and Mastercard are regulated by laws and cannot abuse their dominant position as easily as a large cryptominer can. If Visa or Mastercard began trying to force huge fees on merchants, they’d face fights from very large companies with significant influence — cryptominers would only face fights from individual users, who are not organised and can present much less of a united or influential front against a cryptominer who abuses their position.

If you want a real-world example of this in action, look at the current fight between Amazon and Visa. Amazon is one of few companies big enough to try and fight Visa on its own, but it isn’t doing it out of a genuine belief that Visa charges too much — it just knows it can throw its weight around for its own gain. What is clearly visible, however, is that Visa and Mastercard aren’t invincible; they still have to provide a competitive service due to the ability of large companies to push back on them.