Developer job postings are rising. That still leaves AI’s hiring effect unmeasured

Rising developer job postings leave an important question unanswered: how much hiring, if any, is attributable to AI? A count of advertisements can describe demand without identifying its cause. That distinction matters to a manager deciding staffing levels and to a developer assessing an opportunity. Neither should treat a rebound in listings as a measurement of AI’s employment effect.

CNBC’s October 11 report, citing Indeed Hiring Lab, says U.S. software-development postings rose nearly 15% since late February 2025 while overall postings fell about 7%. It reports that senior roles accounted for 71% of the increase between May 2025 and May 2026, and roles with AI in the title for 37%. These categories can overlap; adding the percentages would be misleading.

CompTIA’s October 2 release reports 47,598 new software developer and engineer postings in September, up 8.3% from August. It also reports a decline of 10,350 technology-sector jobs and 6,000 technology-occupation jobs across industries. Its analysis draws on Lightcast postings and government employment data, with employment estimates subject to revision. These are U.S. figures, not a global hiring measure.

Analysis: compare measures before explaining them

The two reports do not offer one interchangeable score. A monthly count of newly advertised vacancies differs from a change in a posting series over a longer interval. Employment is a third measure. A reader can accept each reported result without assuming that the series share a denominator, population or time window.

Our proposed first step is to write those definitions alongside every number used in a staffing discussion. Record the geography, occupation group, start and end dates, and whether the measure concerns advertisements or employed people. If any definition is unavailable, mark it unknown. This would prevent an apparent comparison from depending on an unstated assumption.

For example, a team should not place September’s developer postings next to a broader technology-sector employment change and subtract one from the other. The occupation and sector boundaries differ, and the units differ. Such arithmetic would produce a result that neither source measures. A useful chart could show the series separately with their definitions visible.

Follow a vacancy through the hiring process

For an employer assessing its own experience, we propose a vacancy register with an opening date, role description, approved staffing purpose and final outcome. Distinguish a replacement from an additional position. Record whether the vacancy was filled, withdrawn or still open. These are suggested internal checks; we have not tested them against the companies in the reports.

That register could help explain what the employer’s advertisements actually meant. A listing that remains open and a person who joins payroll are different outcomes. If the team only has posting data, the completed-hiring result remains unknown. The absence of that information should not become an assumption that every advertisement represents either a new hire or a failed search.

The register should also distinguish a change in wording from a change in work. A proposed review could compare a role’s responsibilities before and after an AI tool was introduced. Adding AI to a title is a different observation from assigning new duties, changing staffing approval or filling a vacancy. Each needs its own evidence.

Do not turn role mix into a skills prescription

The reported senior and AI-title shares describe contributions to an increase over a particular interval. They do not tell an individual applicant which skill will secure a job. Nor do they establish that the same distribution applies to every employer. A candidate can use the report as a reason to inspect actual requirements while keeping that limitation explicit.

Our suggested approach for a job seeker is to compare a small set of current advertisements for the same occupation and location. Note required experience and described tasks separately from marketing language about AI. Check whether the requirements recur across employers. This is a research procedure, not a claim that a particular training course or tool will improve hiring chances.

For a manager, an AI staffing claim would require more than a favorable market headline. A proposed evaluation could document when a tool was introduced, which tasks changed and which hiring decisions followed. Other explanations should remain in view, including changes in projects and staffing budgets. Without a credible comparison, a sequence of events does not identify causation.

What the evidence can support now

The narrower conclusion is that the cited U.S. posting measures show increases, while CompTIA reports declines in broader employment measures for September. That combination does not settle whether AI creates or removes developer jobs. It gives readers a reason to examine definitions and hiring outcomes before drawing a larger conclusion. Future observations should use comparable windows and preserve uncertainty about causes.

Before using a posting rebound to justify an AI staffing decision, compare advertised roles with completed hires over the same period and document what remains unknown.