⚛️Pay & Market

Most physicists specialize early into theory, experiment or computation, and spend their days running calculations, analyzing data, writing code or maintaining apparatus rather than making dramatic discoveries. A large share of the job is incremental: checking a result three different ways, writing it up precisely, and defending it to skeptical colleagues before it counts as established.

For research roles, academic or industrial, yes — a PhD is the field's real credential, since almost no employer trusts someone to design original research without one. A bachelor's or master's degree in physics is still valuable outside pure research, opening engineering, data science, finance and teaching roles that use physics training without the title.

Yes. Physics PhD programs in the US alone graduate roughly 1,800 people a year, while tenure-track openings number in the low hundreds; most physics PhDs never hold a permanent academic position. The training is genuinely excellent for analytical work generally, which is why so many physicists end up well paid in finance, tech or industry instead.

It varies enormously by sector and country. US graduate stipends run around $30,000 a year and postdocs around $60,000, while the median physicist across all sectors earns roughly $150,000 in the US. Physics PhDs who move into quantitative finance can earn several times that, while academic salaries in most other countries are far more modest.

Theorists build and refine the mathematical models that describe how the universe should behave, working mostly with calculations, computers and chalkboards. Experimentalists design and run the apparatus — particle detectors, telescopes, cryostats — that test whether those models actually hold up against reality. Most physicists specialize firmly in one track early in graduate school.

Parts of it. Literature search, first-draft coding, routine data cleaning and figure generation are increasingly AI-assisted. What resists automation is deciding which question is worth years of a career, designing an experiment that can actually answer it, and taking personal responsibility for a claimed result — judgment calls current AI systems cannot be held accountable for.