Physics graduate students in the USA usually also work for their university as teaching assistants or research assistants. The university pays them enough to cover tuition and fees, and live a minimal lifestyle. When I was a graduate student thirty years ago, I shared an apartment with another grad student, did not own a car, and did not have an expensive social life. That left me enough money to buy a few records every month, travel to visit my parents a few times a year, and even visit Europe a couple of times.
How much money do you have to spend doing a Phd (physics)?
Robert, assuming you're talking about tuition, typically a physics PhD costs absolutely nothing, at least in terms of money. Virtually every PhD program in America (and I would guess Europe as well) actually pays you to get your PhD. When you gain admission to a PhD program, most schools will give you a full tuition scholarship with the graduate college. They'll also appoint you as a TA for your first two years, during which you'll get paid something on the order of $15,000 for nine months of teaching. You'll also get paid for your summer research. After your first two years, you'll be appointed as a research assistant. I've seen universities where the graduate college doesn't give you a tuition scholarship, but those schools pay their TAs a higher salary so as to fully offset it. I guess that tuition scholarship vs. additional TA pay is just a question of paperwork. I don't know a single physics PhD student who has ever had to pay his own tuition.
Of course as I hinted above, money isn't the only thing to consider. As those on this forum who've completed their PhDs are fond of saying, it's an extremely arduous path, and you've got to work pretty hard. But if money is the only issue, then rest assured that you won't pay a penny to complete your PhD. Which is somewhat of a relief to people like me, who are just starting their PhDs.
Note that not all schools support all of their physics graduate students. While schools such as Stanford automatically assumes that you will need assistantship when you apply for admission, other smaller schools (especially if they don't have huge endowments) do not automatically give you that, especially before you pass your qualifier. So there's a chance that you may have to pay for your first or second year of grad school, especially if you're an international student.
But after your qualifier, practically every physics graduate students are supported by some form of assistantship. So your 5 to 6 years of Ph.D program may cost significantly less (or none) when compared to your undergraduate cost.
After I worked on my dissertation for about 3 years, the govt agency who funded the work decided to not renew the contract any further so I went out and got part-time jobs and worked on finishing my research on my nickle. After working in industry for a while, they let me go because of budget cuts and I was hurting financially. My saving grace was that the department chair was short of TA's and I spent the final year of my time in grad school teaching freshman lab.
All in all, I paid for about 1/3 of my graduate education out of my pocket.
Wow, I admit that I'm extremely surprised to hear that there are people who've had to pay for grad school. All the programs I applied to said that they were very well-funded, and that all of their grad students were supported by teaching or research assistantships. Have things gotten better in the past few years, or something?