All of the dozens of US STEM PhDs I know from recent decades had to pay their tuition by working as TAs, RAs, or usually both. They didn't walk in the door with a trust fund, work through the coursework, and then write up a dissertation. I'm not sure that's even a possibility in most STEM fields: you need access to millions of dollars of ultra-specialized equipment which you can realistically only get by working in an existing lab, which means being an RA.
Right, they work a job to pay for their tuition. But they still pay, even if it's bartered.
You have a point!