There are no entry-level roles.
That is what I thought. That is what most career services leaders I talk to believe. The narrative has been loud and consistent: AI is automating junior work, companies are not replacing those roles, and a generation of new graduates is entering a market that no longer has room for them.
The data says we are wrong.
Quick Summary
The University of Maryland and LinkUp analyzed 155 million job postings from 2018 through 2025. Entry-level postings are at their highest share in eight years. 12.6% of all 2025 postings explicitly targeted fresh graduates — 63% more than in 2018. The sectors doing the most AI hiring are also the ones adding the most new-grad roles. The data simply does not support the doom narrative. The problem is not that the jobs disappeared. The problem is that students cannot find them.
What the research actually shows
The University of Maryland and LinkUp released their fourth AI Maps white paper in April 2026. They analyzed 155 million job postings across the US from 2018 through 2025. Their goal was to understand how AI adoption was reshaping demand for new graduates.
What they found should reframe how every career services team talks to students right now.
Entry-level job postings are at an eight-year high as a share of all postings. 12.6% of all 2025 job postings explicitly targeted fresh graduates. That is 63% more entry-level postings than in 2018 — the year most people think of as a normal, pre-disruption baseline.
The sectors with the most AI hiring — finance, information services, retail, professional services — are not replacing new grads with AI. They are posting more roles for them. The fear that AI-heavy industries would stop hiring junior talent is not what the postings show.
The data simply does not support the narrative that entry-level is disappearing. What it does support: the system connecting students to those jobs is broken.
Why the doom narrative keeps winning
If the jobs are there, why does it feel like they are not?
Three things are distorting the picture.
The ZIRP baseline problem
The zero-interest-rate era of 2021 produced a historically unusual labor market. Companies were flush with cheap capital and hired aggressively across all levels. That was not normal. It was an anomaly. But students who graduated during that window — and career services professionals who managed through it — established 2021 as their mental baseline for what the job market should look like. When the market normalized, it felt like a collapse. It was not. It was a return to something closer to historical averages.
AI-washing in the media
Layoffs generate headlines. New job postings do not. When 500 junior analysts are let go at a bank, that is a story. When the same bank posts 600 new entry-level roles across its analytics and compliance functions, nobody covers it. Media incentives favor dramatic reversals. The slow, steady accumulation of entry-level demand does not make the feed.
The Handshake bubble
Handshake is convenient. It is campus-integrated. And it captures a fraction of the actual entry-level market. Small and mid-sized employers — who account for a large share of new-grad hiring — post on Indeed, LinkedIn, industry association boards, and direct company career pages. Students who have never been taught to search beyond Handshake are operating with an incomplete view of available roles. They see a small slice of the market and conclude the market is small.
The four real problems
What the UMD and LinkUp data surfaces is not a shortage of entry-level jobs. It surfaces four structural failures in how students are being prepared and supported.
1. The jobs exist but students cannot find them
The entry-level market is not indexed in the platforms students are using. 12.6% of all 2025 postings targeted fresh graduates — but a large share of those postings live outside Handshake and similar campus platforms. This is a search infrastructure problem. Students are not failing because the market failed them. They are failing to access a market they are not being shown.
2. Skills mismatch is real and growing
NACE Career Readiness data for 2026 shows 90% of employers rank problem-solving as a must-have. Students overestimate their own proficiency in leadership and professionalism by 30% or more. Skills-based hiring now covers 70% of employers — up from 65% the prior year. Students are playing a credential game while employers are running a skills-first evaluation. The gap between what students think they can do and what employers are measuring is where most offers are getting lost.
3. Job search is a skill nobody is teaching
Career centers teach resume formatting. They run mock interviews. But the mechanics of actual job search — how to identify relevant opportunities across the full market, how to research employers outside the Fortune 500, how to navigate industry-specific boards, how to calibrate volume and targeting — are not systematically taught. Students graduate with strong credentials and no search intelligence. Then they apply to 20 roles on Handshake, hear nothing, and conclude the market is closed.
4. Expectations were never recalibrated after 2021
Career services teams that managed the ZIRP boom are still working from that playbook. The market normalized. The guidance has not. Students receive framing calibrated to a labor market that no longer exists, which creates anxiety when results do not match expectations — and which fails to equip them for the market that actually does.
What career services can do about it
This is not a headcount problem. Career services teams cannot hire their way to better outcomes. But three things they can do right now would move the needle significantly.
- Expand platform literacy. Teach students to search LinkUp, Indeed, and industry-specific boards — not just Handshake. Show them where the full entry-level market actually lives. That alone changes the search volume and quality of applications.
- Run skills gap assessments against real employer frameworks. Use NACE competency data — not self-reported student confidence — as the benchmark. Close the gap between what students think they can offer and what employers are actually evaluating.
- Make job search a taught skill, not an assumed one. Students who graduate knowing how to find and target the right roles land faster. That is not a given — it is a deliverable. Career services should treat it that way.
The entry-level market is not broken. The system connecting students to that market is. And career services is exactly the function that can fix it — if it is equipped to do so.
Prentus can solve the job search side of this. Aligning academic programs with employer demand is a bigger project and a worthy one. But students who know how to find the jobs that exist, who can demonstrate the skills employers are evaluating, and who are not operating from a distorted 2021 baseline — those students will land. The question is whether your institution is giving them what they need to get there.
