It's certainly a scale thing. What's notable about Amazon is they can run a losslead for an entire sector of industry, not just a single product. Note a grocery store could loss lead toilet paper so much no one can afford to produce toilet paper (maybe); but, could you imagine one which could afford to loss lead _all paper products_ to sell more coffee for long enough to wipe out all other paper retailers?
Because amazon can, and is, doing this systematically. And after that we no longer have fair-market prices, because there is no market but amazons for these goods.
Both exhibit the same behaviors with the same explicit goal.
You are debating semantics when the intent in both cases is the same.
In new markets we tolerate monopolies because we want to reward entrepreneurs.
The question to ask is whether the cost of rewarding Amazon with monopolies is beginning to outweigh the benefits.
I don't think this is the case. The strategy is called a "loss leader", and it's a very common business strategy.
https://www.investopedia.com/terms/l/lossleader.asp
> Loss leading can be a successful strategy if executed properly. A classic example is razor blades. Gillette, for example, gives their razor units away for free knowing that customers must buy their replacement blades, which is where the company makes its profit.
It's true that big companies can find profitable loss-leaders where small companies can't -- though this isn't always true; my startup uses a loss-leader B2B product to unlock a more attractive B2B2C offering.
I think what you're observing is that monopolists tend to be more able to use loss-leaders, since their monopoly power gives them the ability to charge a premium on their monopolized good(s), which would increase the scope of profitable loss-leaders. However loss-leading is certainly not a monopoly-only strategy.