Remote Work May Be Doing More Damage to Young Workers Than AI
Sign Up For Daily Job Market News
Economists Natalia Emanuel (NY Fed), Emma Harrington (UVA), and Amanda Pallais (Harvard) analyzed unemployment data across age groups and occupation types. Their finding: remote work explains 64% of the rise in unemployment among young college graduates since the pandemic. And the timing matters: the surge in youth unemployment began before generative AI spread widely, and the age gap persists even when controlling for AI exposure. “The timing of this surge suggests that remote work, not generative AI, explains the bulk of the rise in youth unemployment,” the authors write.
The numbers are stark. Unemployment for college graduates under 29 averaged 3.1% in 2017-19. It rose 20% to 3.7% in 2022-25. Over the same period, unemployment among more experienced college graduates dipped from 1.9% to 1.8%. Nearly all of the pain landed on younger workers.
The researchers split jobs into remotable occupations (like software engineering) and non-remotable ones (like mechanical engineering). In remotable jobs, young workers’ unemployment rose by almost a full percentage point while older workers in the same fields held steady or improved. In non-remotable jobs, young graduates had a brief bump in 2020 and then went right back to baseline. Remote work has roughly quadrupled since the pandemic, and young graduates are disproportionately in exactly the kinds of jobs that went remote.
The why comes from proprietary data at a Fortune 500 company. When employees work next to each other, they get more feedback and mentorship. Move them apart, even a short distance, and that coaching drops off sharply. The drop is largest for the youngest workers. When the company’s offices were closed, it hired fewer inexperienced workers and more experienced ones who needed less hand-holding. When offices reopened, it went back to hiring junior talent, except on distributed teams, where it kept favoring experienced hires. As the researchers put it, the firm “is willing to teach junior workers when proximity is feasible but shies away from employing inexperienced workers if distance creates barriers to training.”
I’ve seen this from the staffing side for years. Junior hires don’t learn the job from a manual. They learn it from the senior person next to them catching a mistake in real time, from overhearing how a hard conversation gets handled, from the quick redirect that never makes it into a Slack message. Strip that away, and a 23-year-old in a remote setup becomes a training liability rather than an investment. So companies quietly stop making the investment, and the entry-level market gets harder even when openings on paper look fine.
April JOLTS: Openings Jumped, Hiring Fell
The BLS released the April 2026 Job Openings and Labor Turnover Survey this morning, and the split between openings and hiring is the whole story. Job openings rose to 7.6 million, up 731,000 from March, with the rate climbing to 4.6%. Openings are up 520,000 over the past year. Nearly the entire monthly gain came from professional and business services, which added 668,000 openings. Finance and insurance moved the other direction, losing 135,000.
But hiring fell. Hires dropped to 5.1 million, down 419,000, with the rate slipping to 3.2%. The decline was broad rather than concentrated: hiring softened across industries without any single sector driving it. That’s employers posting jobs and not pulling the trigger on filling them, which has been the defining pattern of this labor market for most of 2026.
Quits held flat at 3.0 million and a 1.9% rate. At that level, workers are staying put rather than betting on finding something better. The quits rate is a confidence gauge, and 1.9% tells you workers don’t feel like the door will be open again if they walk through it. That dynamic has been consistent since late 2024. Layoffs were little changed at 1.7 million and a 1.1% rate, retail trade shedding fewer. Still a low-fire market, just a slow-hire one too.
March was revised up to 1.9 million on layoffs and down to 3.2 million on quits. The net picture from revisions: the prior month’s labor market was a bit worse on layoffs and a bit less confident than originally reported.
The connection to the NY Fed paper lands here. Openings exist on paper. Converting them into actual hires, especially for younger and less experienced workers, is where the system keeps stalling. For anyone actively searching for work, the openings are real; the hesitation to move is on the employer side.
Staffing your team doesn’t have to be hard.
Reach out and see how we can help.
Data Centers Promise Thousands of Jobs, But the Permanent Number Is Much Smaller
Communities across the country are being pitched data center projects as economic engines. Lightcast’s new whitepaper runs the actual workforce numbers, and the picture is more nuanced than the press releases suggest.
Building a data center takes a lot of people, somewhere between 1,000 and 10,000 workers, depending on the project. Running one takes almost nobody by comparison, usually 50 to 400 permanent staff. The big job number communities hear at the ribbon-cutting is mostly temporary construction work that ends when the building is done.
Lightcast modeled this in Laredo, Texas. A thousand construction jobs pump roughly $73 million in earnings into the local economy and support 1,443 total jobs through ripple effects. But 100 permanent operational jobs generate only $8.2 million and 174 total jobs, roughly matching what house builders ($8.0 million, 148 jobs) and machinery repair ($7.9 million, 157 jobs) would produce from the same labor pool. Their conclusion: the ROI on permanent data center jobs is “good but not exceptional.”
Then there’s the labor supply problem. The construction phase draws on electricians, HVAC techs, equipment operators, plumbers, and inspectors. Lightcast pegs the skilled trades shortfall at 1.7 million workers a year nationally: 2.9 million annual openings against only 1.25 million qualified graduates. Data center construction postings hit 24,721 over six months, up 23%, and have roughly doubled in two years. These projects are all competing for the same limited pool of workers at the same time, in every region of the country at once. Nick Lee Romagnolo of Amazon Web Services put it plainly: “What used to happen is we would bring in travelers from other places to help meet those needs, but now these needs are everywhere across the US.”
The operations side has its own bottleneck. Demand for data center technicians and engineers ran 27,650 postings over six months, up 23%, from more than 4,400 companies. The skills gap is sharp: “Data Centers” as a skill shows up in 32% of job postings and only 7% of worker profiles. Remote data center tech roles climbed to 10.2% of postings in 2025, up from 6.7% in 2022, meaning even the few permanent jobs may go to workers who don’t live in the community being promised the economic benefit.
For mayors and county commissioners being pitched on data center development: the construction spend is real, the tax revenue is real, and the community college investment to build the trades pipeline is a genuine long-term win. But “thousands of jobs” is mostly a construction crew that leaves when the building is done. The workers left behind will face a familiar challenge: demand that exists, and a pipeline that can’t move fast enough to meet it.
