Those customers are called Deadbeats[1]. All credit card customers should strive to be Deadbeats.
* Amex had 67b in interest + non-interest revenue, of which the single biggest driver was discount revenue (aka interchange) at 33b. Interest contributed 20b, card membership fees contributed 7b, and services fees and other were ~7b combined. Amex's target base is wealthier individuals.
* Discover had 14.4b in credit card interest revenue and 1.4b in net discount and interchange revenue (meaning after paying rewards). Discover caters mostly to what the financial industry would euphemistically call "subprime" creditors.
On top tier cards, interest and late fees are not the intended way to make money on the majority of cardholders (although of course banks will still charge them for customers that do carry a balance). On lower tier cards, you'll often see less rewards points or even annual fees without any corresponding benefits (i.e. no lounge access, no rental car insurance etc).
For example, American Express has traditionally focused on higher-income cardholders, and their top tier cards are charge cards, which are always due in full at the end of each statement period (i.e. carrying a balance is not an option). They don't even list an APR, as far as I remember (although they've recently started offering an option to carry certain purchases, making that story a bit less clear).
I presume that the type of person who wants an apple card is strongly negatively correlated with the kind of person from whom they could make lots of money via fees from. I would bet my apple card balance that the percentage of apple card customers who religiously pay off their balance every month is much much higher than industry average.
GS did have some other cards before Apple Card, but they were not generally marketed or widely available to the masses like Apple Card. GS wasn’t well prepared for the extra overhead and customer support, and Apple negotiated a _lot_ of power that GS agreed to give them in areas like underwriting decisions. Apple wanted applications approved and many approvals wouldn’t have met with GS’s typical risk tolerance.