AI job fears collide with a hiring market that is splitting, not collapsing

On 25 September, CNBC reported that recent college graduates are entering a job market that is tougher than the headline numbers suggest.

According to the Federal Reserve Bank of New York, graduates aged 22 to 27 had an unemployment rate of 5.7% as of June 2026. That is a full percentage point higher than two years earlier for the same cohort. Underemployment, meaning graduates working in jobs that do not require a bachelor's degree, reached 42% for recent graduates, against 33.7% for all college graduates. That is the anxiety side of the story.

The next day, a different CNBC report, published on 26 September, described something close to the opposite. Welders, pipefitters, electricians and HVAC technicians are in demand for data centre construction, and pay is moving fast. Nicole Bachaud, a labour economist at ZipRecruiter, told CNBC that the mean minimum salary for data centre jobs jumped 125.1% year over year to nearly $208,000. Postings for welders and pipefitters were up 164% year over year.

Two labour markets, one AI boom

The two reports do not cancel each other out. They describe different segments of the same transition, and those segments are not equally visible in aggregate data. Maria Flynn, president and CEO of the nonprofit Jobs for the Future, told CNBC that data centres are only one piece of a larger puzzle, with grid upgrades and transport modernisation adding demand at the same time. "We are seeing major sources of demand that are converging at the same time," she said. Justin Sinkovich, an associate professor at Columbia College Chicago, put specific numbers on that demand for CNBC.

He pointed to Louisiana, where Amazon committed $12 billion to a new data centre with 540 on-site jobs, including 1,700 electricians, technicians and security personnel. Meta's Hyperion project in Louisiana, he said, is worth $27 billion. "These projects create demand for on-site workers and therefore cannot be offshored or conducted remotely, unlike manufacturing and software," Sinkovich told the network.

Not every region gets the work. Bachaud said Houston and Birmingham have seen particularly strong growth, while coastal hubs face higher costs and tighter regulation. Her read: an electrician in Birmingham will probably find a better market than one in Boston.

The construction jobs, however, are largely temporary. CNBC noted that many of the buildout roles end when the building does. That leaves the question of what happens to those workers once the data centre is running.

Meanwhile, the same technology is reshaping the entry-level white-collar market. CNBC's 25 September piece cited LinkedIn research showing that 66% of recruiters planned at the start of 2026 to increase their use of AI for pre-screening interviews, while 81% of job seekers said they had used or planned to use AI in their search. Handshake, an early-career platform that has worked with OpenAI, found that graduating seniors in 2026 mentioned AI skills on their resumes at twice the rate of the class of 2022. It also found that 74% of those mentions were tied to real-world projects rather than coursework. The National Association of Colleges and Employers said AI skills appeared in 16.5% of job descriptions in the spring, up from 10.5% the previous fall. Overall, 28% of employers said they were seeking early-career talent who can use AI.

What the data does not show

The most careful reading of the evidence comes from the New York Fed, and it is less dramatic than either the panic or the reassurance. Based on its study of job postings, the NY Fed wrote that "while AI may be contributing to recent labor market developments, it is not the main driver of the slowdown in hiring." Its survey work with employers found they "mostly intend to incorporate AI mainly via retraining, with limited effects on hiring."

That is a useful corrective to a week of alarming headlines. The dossier also includes a headline from finance.biggo.com dated 26 September: "Fears of AI-Driven Entry-Level Hiring Collapse Not Yet Backed by US Jobs Data." The same cautious framing runs through a Pioneer Press column from 27 September and a comment piece in The News Pakistan the day before. Both argue that the predicted entry-level collapse has not yet shown up in the numbers.

So where does that leave a graduate deciding what to do? The honest answer is that the market is splitting rather than shrinking. Trades tied to physical infrastructure are paying more, and AI is not building substations or cooling systems. Office-track entry roles are being screened harder, and applicants are responding by demonstrating projects rather than listing skills. The Fed's research suggests retraining, not replacement, is the dominant employer plan so far. There is one more variable, and it is political rather than technological.

CNBC's 26 September report noted that states including New York and Texas have moved to slow data centre development, and local freezes have followed public backlash. If those restrictions spread, the blue-collar boom that AI helped create could be the first thing it loses. That would not show up in a graduate unemployment rate. It would show up in welders' paychecks.

None of the sources in this dossier claim to know which way that goes. The New York Fed says AI is not the main driver of the hiring slowdown. ZipRecruiter's data says specialised data centre roles are paying more than ever. Both can be true at once, in a labour market that is no longer moving in one direction.