I suppose for everyone we reach a point when we have to reconcile our dreams with the practical reality of how to make a living.
This is very important -- I have a lot of respect for people who are willing to devote their life to research and advancing human knowledge but that's not a life for me! I don't think a PhD is something one should do just because it's the "logical next step" in one's education. It's ~5 years in a fairly important phase of your life/career development! It's important to set goals and know exactly what you want to accomplish from your PhD.
For me, my goal is a career where I can support my future family while having a job that I enjoy going to. But family is more important to me, and moving around constantly for post-docs is not really ideal. So my wife and I have decided that I'm only going to look for post-docs/permanent jobs in our hometown (maybe a post-doc can be in a really cool place to live) so that we can be close to our family when we raise our kids. I know it's hard enough to get a faculty job without being willing to move anywhere, so while I will still try for one, I am realistic about the low probability. But, my PhD will help me get other jobs in our hometown (it's a big city) at non-research colleges, science centres, etc. I am also perfectly happy with working as full time permanent research staff in a lab or group at a University! Knowing this, I plan to make sure I develop skills during my PhD that are also valuable outside of the research environment, such as teaching and communication.
The one thing that remains confusing is how does someone determine a " safe school". From what I can discern the admittance process is multi-faceted and there is a great many variables as to why an individual is accepted or rejected.
There are hundreds of programs in physics. To thoroughly explore each school and professor would take hundreds of hours - even if you only devoted one hour of time for each one.
Your son's stats look like they are very competitive. He performs strongly in every metric I think -- research, GPA, test scores. But I think the people who will know your son's chances the best are his research advisors. It's extremely helpful, in my experience, to have a mentor that is able to guide you through grad school applications. This person can honestly tell your son where he stands a good chance to get in and where he will almost certainly get in (i.e. safeties). I'm probably the first person in my family to be in grad school so without my mentor (undergrad thesis advisor), I wouldn't have had anyone to ask for advice.
I talked to my mentor and my current supervisor before I even started looking at PhD programs, asking them who they think the good supervisors and schools are. Then, I checked in with them again when I made a shortlist of schools and potential supervisors. I had yet another conversation after getting results and visiting the schools and sharing my impression of each place.
I got 2 really helpful pieces of advice during this process, but this is more helpful in a year from now, when your son has to decide between offers.
1. You should pick a project that will be interesting to people who are going to hire you in ~5 years. Even if you are super passionate about some topic, if no one else is, it's not going to really help you get a job. Your PhD work will be your "calling card" into getting post-docs, and you want to do something interesting that gets your name associated with that topic and vice versa. Post-docs are hired pretty entirely on their ability to do interesting independent research.
2. A great researcher does not equal a good supervisor. A PhD is more than just doing good research -- it's learning how to develop skills to be a fully independent scientist. A good advisor will train their students to give good presentations, write good papers, interact with other scientists effectively as well. Profs in the field may be familiar with other profs' students through conferences etc. and they could get a good sense of how well the prof treats and trains their students. You can also get this information from talking to current students during school visits. The advice I got if I was interested in a great researcher but it turns out they are a very bad PhD advisor was to work with them later as a post-doc, or collaborate with them with another person who is actually a good advisor. Again, when you are looking for post-docs or tenure-track positions, you will be giving a lot of job talks and you're evaluated on how well you perform research as well as how well you present your work.
Like you said, you know about the general process of applying to PhDs but you couldn't know the recent trends in the subfields of Physics, or how well your son really compares to other students, or what kind of students are being accepted into certain programs etc. My mentor was great for helping me out in these questions. So, I know you are trying to help your son, but I really think the best person/people to do so are faculty members that know him well and are willing to mentor him. It could also help them understand your son's interests and goals more strongly and that could turn into a more effective reference letter.
I turned to my parents and family for advice for the non-research aspects of grad school life: how far away from my family am I comfortable living, should I rent an apartment or try to buy a house, when to start a family, determining my budget (grad school was the first time I was truly living on my own), what to do about other adult responsibilities such as car/renters/life insurance, whether or not it's worth it to spend X dollars on a big repair on my old car, investing money/saving for the future, doing taxes, and all that other stuff!!
My parents did not get the chance to go past high school for their education so I wouldn't have been able to turn to them for grad school advice anyways. But unless they were physicists, I probably would still sought this advice from one of my supervisors/mentors, since these people have worked with me and know my strengths and weaknesses while also knowing what's going on in the physics research world.