Entry-level postings are down 35%. New grads are paying the price.

The numbers: a degree no longer beats the average

For decades, the deal was simple: get a degree, and your unemployment odds would be better than the overall workforce's. That deal quietly broke. Recent graduates aged 22 to 27 now face 5.6% unemployment while the overall rate sits at 4.2% — a historic reversal. For the class of 2026 specifically, unemployment reached 5.7%, the worst since the height of the pandemic.

Note the last number. Employers actually project more hiring for the class of 2026 than last year. The problem isn't that companies stopped hiring young people. It's that they stopped hiring young people to do entry-level work.

What's actually disappearing

The roles shrinking fastest read like a list of what large language models do well: junior analyst roles, customer support positions, research assistant gigs, basic coding tasks. These were never glamorous jobs. They were something more important — on-ramps. The place you spent two years learning how the industry works while doing work that was, frankly, automatable.

Now it's being automated. Employers expect new hires to slot in at a higher level almost from day one. The training ground where juniors became mid-levels is being absorbed by tools that don't need training, don't need benefits, and don't need a desk.

This creates a paradox that every 2026 grad knows by heart: entry-level jobs now require experience. And the data shows just how brutal that filter is:

The internship is no longer a nice-to-have. It is the entry-level job now — it just happens before graduation and often pays accordingly.

The part that doesn't fit the headline

If AI were simply deleting white-collar work, you'd expect white-collar employment to shrink. It hasn't. Since ChatGPT arrived in late 2022, the US economy has added about 3 million white-collar jobs. There are 7% more software developers than in 2022, 10% more radiologists, and 21% more paralegals — three professions that were supposed to be first against the wall.

And there's a rival suspect with a strong alibi problem: remote work. Research highlighted by NPR in June 2026 found that companies have become much less likely to hire recent graduates into occupations that can be done remotely. The logic is uncomfortable but plausible: training a 22-year-old over Zoom is hard, mentoring them is harder, and if the work can be done remotely, it can also be done by an experienced contractor three time zones away — or, yes, by software.

The honest answer is that AI and remote work are compounding each other. Remote work made juniors expensive to train. AI made their starter tasks cheap to automate. Either force alone might have dented the entry-level market. Together, they removed the bottom rung of the ladder — and everyone still standing on the upper rungs barely noticed.

Why this matters even if you're not 22

The entry-level job was never just a job. It was the industry's training pipeline. If companies stop developing juniors for five years, the mid-level shortage arrives right on schedule in year six — and seniors who were never replaced by AI will be replaced by simple demographics: retirement, burnout, attrition, with no bench behind them.

It also changes the layoff math for everyone else. When laid-off workers with three years of experience compete for the same roles as new grads, the grads lose — and the experienced workers ratchet down, accepting roles a level below their last one. Underemployment rolls downhill. That's how 43% of recent graduates end up in jobs that don't require a degree they're still paying off.

One more thing: companies announcing AI-driven job cuts often have other reasons — over-hiring, offshoring, plain cost-cutting — and "AI" makes a better press release than "we misjudged headcount." We track which layoffs are genuinely AI-attributed, with receipts, on our layoffs tracker.