Technology employers and technical work
Examples use releases available September 5, 2026. Dates and survey coverage are stated beside the findings.
Technology is not a single employer industry or occupation. Comparing several industries, in both job counts and percentage changes, helps reveal patterns that a single headline can hide.
Which technology industries grew?
Between January 2019 and July 2026, seasonally adjusted CES payroll employment changed as follows. Counts are thousands of jobs; July 2026 estimates and derived changes are preliminary.
| Employer industry | January 2019 | July 2026 | Net change | Percentage change |
|---|---|---|---|---|
| Software publishers | 440.9 | 657.0 | +216.1 | +49.0% |
| Computing infrastructure, data processing and web hosting | 330.9 | 461.0 | +130.1 | +39.3% |
| Computer systems design and related services | 2,153.6 | 2,363.6 | +210.0 | +9.8% |
| Semiconductor and other electronic component manufacturing | 374.1 | 369.6 | −4.5 | −1.2% |
| Computer and peripheral equipment manufacturing | 123.5 | 99.8 | −23.7 | −19.2% |
Sources: software publishers, computing infrastructure, computer systems design, semiconductor and electronic components, and computer equipment.
Software publishers had both the largest percentage increase and the largest job gain among these five industries. Computer systems design remained much larger in employment level. The comparison illustrates why a growth ranking and an employer-size ranking can tell different stories.
An industry is not an occupation
A programmer working for a bank belongs to banking in an industry comparison. CPS occupation estimates instead group people by their primary work. CES payroll jobs and CPS employed people are different measures and cannot be added or used to allocate one another's gains. Detailed CPS occupations generally require same-month annual comparisons, with attention to seasonal adjustment, sampling variability and January population-control changes. Sources: CES FAQ and CPS definitions.
Does this show the effect of AI or subsidies?
No causal effect is identified by these employment paths alone. For example, software-publisher payrolls rose by 15,100 jobs, or 2.4%, from January 2024 to July 2026, seasonally adjusted. The July endpoint is preliminary. That measured change does not establish whether AI caused it. Source: software-publisher payrolls.
Attributing a change to AI adoption or a subsidy would require information about exposure, timing and a defensible comparison group. Those inputs are not available in these national employment series. Similarly, broad JOLTS industry openings do not reveal how many jobs target software engineers, recent graduates or other specific occupations.
Continue with work and opportunity or compare employment series.