The Bureau of Labor Statistics released its Occupational Employment and Wage Statistics on 15 May 2026, covering the May 2025 reference period. It is the most authoritative national wage table available, built from an employer survey rather than self-reported salaries, and it is the correct source for anyone comparing STEM pay across fields.
One naming point first, because it is a persistent source of confusion: this is the May 2025 dataset, published in 2026. Coverage that labels it "2026 wage data" is describing the release year, not the period the wages were measured in. Any comparison against an earlier release has to line up reference periods, not publication dates.
Read against the all-occupation national mean annual wage of $69,770, the STEM premium is real and large. But the table also contains a pattern that most "highest-paying careers" lists omit entirely, and it changes what a student should do with the ranking.
Where STEM Pay Sits by Field
At the major-group level, the ordering is consistent and unsurprising:
| Occupational group | Mean annual wage | Against the national mean |
|---|---|---|
| Computer and mathematical occupations | $120,080 | +$50,310 |
| Architecture and engineering occupations | $106,830 | +$37,060 |
| Life, physical, and social science occupations | $94,600 | +$24,830 |
| All occupations | $69,770 | — |
Computing leads engineering by roughly $13,000 at the group mean, and engineering leads the sciences by roughly $12,000. That ordering holds at the level of individual occupations too, but the group averages conceal a spread wide enough that field alone is a weak predictor of pay.
The Ranking, With the Column That Usually Gets Cut
Below are the highest-paying STEM occupations in the May 2025 data, with national employment shown alongside the wage. The second column is the one that changes the interpretation.
| Occupation | Mean annual wage | National employment |
|---|---|---|
| Computer and information systems managers | $192,160 | 670,570 |
| Architectural and engineering managers | $181,540 | 220,260 |
| Physicists | $171,180 | 20,430 |
| Computer hardware engineers | $162,670 | 76,660 |
| Petroleum engineers | $157,150 | 18,060 |
| Computer and information research scientists | $153,930 | 37,200 |
| Software developers | $148,100 | 1,687,890 |
| Aerospace engineers | $142,060 | 67,710 |
| Actuaries | $141,480 | 26,670 |
| Nuclear engineers | $140,420 | 15,280 |
| Computer network architects | $139,580 | 179,740 |
| Astronomers | $132,920 | 2,120 |
| Information security analysts | $132,510 | 190,650 |
| Chemical engineers | $129,980 | 21,070 |
| Biochemists and biophysicists | $129,840 | 33,830 |
| Mathematicians | $129,260 | 2,030 |
| Data scientists | $126,800 | 262,440 |
| Electrical engineers | $125,100 | 198,750 |
| Statisticians | $115,700 | 29,030 |
| Mechanical engineers | $113,610 | 296,810 |
| Civil engineers | $108,670 | 367,840 |
| Epidemiologists | $97,890 | 12,090 |
High Pay and Small Occupations Travel Together
Astronomers earn a mean of $132,920 across 2,120 jobs in the entire United States. Mathematicians earn $129,260 across 2,030. Nuclear engineers earn $140,420 across 15,280, and petroleum engineers $157,150 across 18,060. These are genuinely excellent wages attached to occupations small enough that a single metropolitan area may hold a few dozen positions.
Software developers, by contrast, earn $148,100 — above every engineering discipline in the table, and seventh overall — across 1,687,890 jobs. That is roughly eight hundred times the size of the mathematician occupation. Data scientists at $126,800 hold 262,440 positions, and information security analysts at $132,510 hold 190,650.
So the practical reading of this table is not the ordering of the first column. It is which rows combine a high wage with enough positions that the wage is reachable through ordinary means — a normal degree, a normal job search, in a normal labour market. On that measure computing dominates, not because it pays the most, but because it is the only STEM area where top-decile pay exists at genuine scale.
This is the same distinction we drew in examining the fastest-growing STEM occupations, where headline growth rates attached to small occupations produced fewer annual openings than a slow-growing large one. Rate and rank both mislead in the same direction, and for the same reason: neither is a measure of how many jobs there are.
The Two Rows at the Top Are Not Entry Points
Computer and information systems managers ($192,160) and architectural and engineering managers ($181,540) lead the table, and both are management occupations that BLS classifies outside the technical groups. They are reached after years in a technical role, not applied to from a degree programme.
That is worth separating out, because their presence at the top of every "highest-paying STEM jobs" list creates a misleading impression of what a technical career pays. A more useful way to read those two rows is as the ceiling on a computing or engineering track rather than as an alternative to one. The relevant comparison for a student choosing a major is the highest-paying individual contributor row — physicists at $171,180 for those going the doctoral research route, computer hardware engineers at $162,670, software developers at $148,100 at scale.
What Mean Wages Do and Do Not Tell You
The figures above are means, and means in wage distributions sit above medians because high earners pull them upward. For software developers the gap is material: the mean of $148,100 runs well above the median, so the average is not the wage a typical entrant should expect. Anyone using these numbers for planning should pull the median and the percentile breakdown for the specific occupation, both of which OEWS publishes.
Two further limits are worth stating. These are national figures, and geographic variation within a single occupation frequently exceeds the variation between occupations — a software developer's pay differs more across metropolitan areas than the gap between software development and hardware engineering nationally. And OEWS measures wages for people already employed in an occupation; it says nothing about the probability of entering it. A high mean wage in a small occupation with few openings is a description of the people who got in, not a forecast for someone trying to.
The most defensible use of this release is comparative and directional: STEM pays substantially above the $69,770 all-occupation mean, computing pays the most among the large fields, and the fields with the highest headline wages are mostly too small to plan a career around on the strength of the wage alone.
Is this the newest federal wage data available?
Yes. The May 2025 OEWS estimates were released on 15 May 2026 and are the current national wage table. OEWS runs on an annual cycle with roughly a one-year lag between the reference period and publication, so a gap of that size is normal rather than a sign of stale data.
Why is software development ranked below physics if it pays better in practice?
It is ranked by mean annual wage, and physicists have the higher mean. What differs is scale and accessibility: 20,430 physicist positions, most requiring a doctorate, against 1,687,890 software developer positions with far more varied entry routes. The ranking answers "which occupation pays more," which is not the question most people are actually asking.
Should I choose petroleum or nuclear engineering for the salary?
Both pay exceptionally well and both are small — 18,060 and 15,280 positions nationally — with employment concentrated in particular regions and sectors. That concentration is the risk: these occupations are more exposed to industry-specific and policy-driven demand shifts than a large, sector-diversified occupation. The wage is real; the assumption that a job will be available where you want to live is the part to check.
Does a STEM degree guarantee a wage above the national mean?
No. Occupational means describe occupations, not graduates, and a STEM degree does not guarantee employment in a STEM occupation. The premium in this table is attached to working in the occupation, which is a separate outcome from completing the credential.
Where do the sciences sit relative to computing and engineering?
Lower at the group mean — $94,600 for life, physical, and social science occupations against $120,080 for computing and $106,830 for engineering — but with high-paying exceptions inside the group, including physicists at $171,180 and biochemists and biophysicists at $129,840. Epidemiologists at $97,890 sit near the group average despite the doctoral training common in the field.
← Back to Careers | AI and Machine Learning Careers | STEM Careers Without a Degree →