You've got the sequence of motivations and expectations of investments & revenue backwards.
A more accurate timeline is that before the product is built, the businesses will segment the potential markets. They estimate the various price curves of price-sensitive and price-insensitive customers and then use that to determine how much to allocate to each feature. Some features are valuable to price-insensitive customers. Some features are useful to all customers.
To say the above was "investment already made" is flawed because that perspective depends on reversing the timeline around in a way that businesses don't actually think. Instead of "sunk cost", it's actually a "before-we-build-it calculated investment" based on "future revenue projections".
Likewise, when a pre-med student considers going to medical school, he may envision a future where he provides medical care for free in Africa and at the same time earn a $200k+ salary as a hospital surgeon. Those future market segments affects his current decision to spend (go into debt) $150k+ for his medical degree. When the doctor is 40-something, we, as patients can't say that he should give us medical care for minimum wage because we morally deserve his services and his medical school tuition 10 years ago is a "sunk cost" anyway.
As for how a business delivers that segmentation of product to a customer, that is a separate issue. Some methods are perceived as more distasteful than others. The tig welder with a software config switch causes more outrage than a different physical chassis.