Skip to main content
  • Home
  • News - U.S.
  • College Grads Face Hiring Chill as Employers Battle for High School Grads: AI Triggers a ‘Degree Reversal’ in the U.S. Job Market

College Grads Face Hiring Chill as Employers Battle for High School Grads: AI Triggers a ‘Degree Reversal’ in the U.S. Job Market

Picture

Member for

1 year 9 months
Real name
Lauren Robinson
Bio
Vice Chief Editor
With a decade of experience in education journalism, Lauren Robinson leads The EduTimes with a sharp editorial eye and a passion for academic integrity. She specializes in higher education policy, admissions trends, and the evolving landscape of online learning. A firm believer in the power of data-driven reporting, she ensures that every story published is both insightful and impactful.

Modified

Generative AI absorbs document drafting, data analysis and basic coding
Entry-level hiring freezes as demand for electricians, plumbers and welders surges
Occupational polarization accelerates, eroding the college-degree premium

A reversal of traditional employment polarization is emerging in the US youth labor market. Young people without college degrees are enjoying their strongest employment conditions in two decades, while the job market for college graduates has deteriorated to its weakest level since the global financial crisis. As the proliferation of artificial intelligence (AI) reduces demand for the entry-level clerical and professional roles typically filled by recent graduates, severe labor shortages have driven up the market value of skilled tradespeople—including electricians, plumbers and welders—needed to build data centers and expand the power grid. The long-standing assumption that a college degree guarantees access to a better job is being challenged by the advent of AI.

Graduate–Non-Graduate Job-Finding Gap Nears Half-Century Low

According to the Burning Glass Institute, a US labor-market research organization, on Sept. 8 (all dates local time), the unemployment rate among 22- to 34-year-olds without college degrees has approached its lowest level in more than two decades. The trend reflects an intensifying shortage of skilled tradespeople amid an aging workforce, rising retirements and declining immigration. Unemployment has been particularly low in occupations involving substantial manual labor and face-to-face work. These include production and on-site jobs in services, construction and manufacturing that are difficult for AI to replace. The August jobs report released by the US Bureau of Labor Statistics (BLS) on Sept. 4 reinforces this trend. Restaurant and bar employment accounted for more than one-third of the 162,000 jobs added in the United States last month.

College graduates, by contrast, are facing their worst employment conditions since the 2007 global financial crisis. The difficulties were found to be particularly acute among job seekers with master’s degrees or higher qualifications and those who majored in science, technology, engineering and mathematics (STEM). “College and non-college job seekers are experiencing diametrically opposed conditions,” said Gad Levanon, chief economist at the Burning Glass Institute. “This contrasts with the customary pattern in which employment conditions for the two groups generally moved in the same direction, irrespective of the business cycle.”

The gap in job-finding rates by educational attainment has also narrowed. A recent Cleveland Federal Reserve analysis of the US Census Bureau’s Current Population Survey (CPS) from 1976 to 2025 found that the unemployment-rate gap between high school and college graduates aged 22 to 27, which had remained at around 5 percentage points for decades, narrowed to a 12-month moving average of 2.5 percentage points in July. This was close to the 2.4-percentage-point gap recorded in March 2024, the lowest level since the late 1970s. The proportion of unemployed people who subsequently found jobs also converged across the two groups around 2019.

Entry-Level White-Collar Hiring Plunges as AI Adoption Expands

The employment divide by educational attainment has widened as the spread of AI drives labor demand in sharply different directions across occupations. In a survey of Texas companies conducted in May, two-thirds of respondents said they were using AI in their operations. The adoption rate had risen from roughly 40% two years earlier to more than 60%. An analysis of millions of online job postings by the Federal Reserve Bank of Dallas found that postings for occupations with a high share of tasks automatable by generative AI were approximately 5% lower than those for low-exposure occupations at the end of 2023 and 8% lower in the first quarter of 2025. Related job postings at companies with high levels of AI usage also declined by 8–9% through early this year.

Companies first cut the entry-level tasks traditionally assigned to new hires. Generative AI can draft documents, summarize materials, conduct standardized analyses, respond to customers and perform basic coding. The high-exposure occupations identified by the Dallas Fed included not only software developers, web designers and computer programmers but also clerical workers, editors and some managers. Given that more than half of online job postings target applicants with no more than two years of experience, the impact of the decline is inevitably concentrated among recent college graduates and workers preparing to change jobs. Researchers estimated that exposure to generative AI reduced total online job postings in Texas by 1.8% in 2024 and 2.6% last year.

Table 1. Changes in Job Postings by Texas Companies Following the Expansion of Generative AI

CategoryPeriodKey FigureDetails
AI adoption rate among Texas companiesMay 2026Approximately two-thirds of respondentsIncreased from approximately 40% two years earlier to more than 60%
Job postings in occupations highly exposed to generative AIQ1 2025Approximately 8% lower than in low-exposure occupationsIncluded software developers, web designers, programmers, clerical workers, editors and some managers
Related job postings at existing companies with high levels of AI usageThrough early 2026Down 8–9%Reductions in entry-level tasks including document drafting, material summarization, standardized analysis, customer service and basic coding
Total online job postings attributable to generative AI exposure2025Down 2.6%More than half of online job postings targeted applicants with no more than two years of experience
Source: Federal Reserve Bank of Dallas

AI Capital Surges While Employment and Income Indicators Deteriorate

Texas is particularly noteworthy because AI infrastructure investment and graduate employment have followed markedly different trajectories. According to global real estate services company Jones Lang LaSalle (JLL), the combined capacity of operational and under-construction data centers in Texas reached 26 gigawatts (GW) in the first half of this year, twice the 13GW recorded in Virginia, the second-ranked state. Texas therefore accounted for a substantial share of the 66GW of data center capacity under construction across North America. Employment and income indicators for Texas college graduates examined by the Dallas Fed, however, deteriorated over the same period. Data center investment and entry-level graduate recruitment did not increase in tandem.

Research has also found that the employment effects of data centers are limited. The Brookings Institution linked data on approximately 1,500 US data centers and 52 canceled projects with BLS statistics to examine changes in regional employment between 2003 and 2024. It estimated that the arrival of a region’s first large data center typically generated only 100–200 additional jobs over a decade. Employment in data processing increased by 56% and telecommunications employment rose by 43%, but the initially small number of workers in both industries limited their effect on overall employment. Local wages did not rise significantly, while home prices increased by 2–5%.

Youth Employment Falls 6% in AI-Exposed Occupations

At the same time, the generative AI that triggered the data center investment boom is absorbing entry-level tasks in the clerical and professional occupations traditionally pursued by college graduates. An analysis of data from US payroll-processing company ADP by Stanford University’s Digital Economy Lab found that employment among 22- to 25-year-olds in highly AI-exposed occupations declined by 6% from late 2022 through September 2025. Employment among workers aged 35 to 49 in the same occupations, by contrast, increased by more than 8%. In software development, employment among 22- to 25-year-olds fell by approximately 20% from its late-2022 peak.

Changes are also emerging in corporate hiring practices. According to Chicago-based recruiting firm Hirewell, advertising and marketing agencies have recently hired almost no entry-level employees. Consulting firm Futurety stopped recruiting interns this summer and subsequently assigned ChatGPT to produce social media posts. As experienced employees and AI jointly handle research, document drafting, basic analysis and content production, companies have less incentive to recruit and train new employees. The employment shock facing college graduates is therefore appearing primarily through reduced recruitment rather than layoffs of existing employees.

Blue-Collar Pay Surges Amid Data Center Construction Boom

Demand for on-site workers, by contrast, has surged during the construction of data centers and expansion of power grids. Operating servers requires high-capacity substations, cooling systems, fiber-optic networks and emergency power-generation facilities. Efforts to secure electricians, plumbers, welders, heating, ventilation and air-conditioning (HVAC) technicians and heavy-equipment operators have consequently spread across major US construction sites. These occupations often do not require four-year college degrees, with apprenticeship completion, professional certifications and on-site experience serving as the principal hiring criteria.

According to labor-market analytics company Lightcast, postings for data center-related construction jobs increased by 23% over the past six months and nearly doubled within two years. Demand for data center technicians and engineers rose at a similar pace. The annual US shortage of skilled tradespeople is already estimated at more than 1.7 million workers. In a representative project model calculated by Lightcast using the Laredo, Texas, area, 1,000 workers were required during the construction phase alone, generating an estimated $74 million in labor income.

Competition for on-site workers has also spread to the power industry. Consulting firm Deloitte projected that electricity demand from US data centers would rise from 47GW in 2025 to more than 176GW in 2035. Between 2023 and 2025, job postings for core data center occupations increased by 64% and those in the power industry rose by 20%, while nationwide postings for the same occupations grew by only 4%. Postings for data center electrical-equipment technicians surged by more than 180%, while demand for power-generation equipment operators increased by 56%. More than one-third of the new postings issued by the two industries targeted the same occupational groups.

Meta and Google Invest Heavily in Skilled-Trades Training

Labor shortages are translating into higher wages and improved working conditions. According to The Wall Street Journal (WSJ), some skilled tradespeople who moved to data center construction sites are earning 25–30% more than in their previous jobs, with a growing number receiving annual salaries exceeding $100,000. A survey by staffing company Kelly found that 90% of data center operators regarded labor shortages as a significant risk to project execution. The national median annual salary was reported at $112,000 for specialized data center technicians and $128,000 for construction managers, while one in four on-site workers changed jobs after being recruited by a competitor.

With access to labor increasingly determining project schedules, major technology companies have begun directly investing in skilled-trades training. Meta committed $115 million to launch America’s Workforce Academy, a free vocational training program operating in Louisiana, Ohio, Indiana and Texas. The program, which guarantees graduates an industry-recognized credential and employment with a partner company, attracted 35,000 applicants within seven days of opening. Google also committed $50 million to train 300,000 skilled workers across more than 20 US states. Eligible programs include training for electricians, welders, plumbers and HVAC technicians.

Picture

Member for

1 year 9 months
Real name
Lauren Robinson
Bio
Vice Chief Editor
With a decade of experience in education journalism, Lauren Robinson leads The EduTimes with a sharp editorial eye and a passion for academic integrity. She specializes in higher education policy, admissions trends, and the evolving landscape of online learning. A firm believer in the power of data-driven reporting, she ensures that every story published is both insightful and impactful.