The headline number looks like good news. Employers project 5.6 percent more new graduate hires for the Class of 2026 compared to last year, according to the National Association of Colleges and Employers’ spring update. 77.2 percent of recent graduates landed a role within three months of graduating, up from 63.3 percent a year earlier. Set those numbers against the 5.6 percent unemployment rate for recent graduates aged 22 to 27 — one of the highest in a decade outside the pandemic — and the picture becomes considerably less straightforward.
What sits underneath all of this is one statistic that makes the rest of the data coherent: graduates who completed an internship or co-op were hired at a rate of 81.6 percent. Graduates without any work experience were hired at 40.7 percent. That 41 percentage point gap is the real story of the Class of 2026.
The market is better than last year and worse than the headlines suggest
The 5.6 percent increase in projected hiring is a spring revision. The original autumn projection, before employers updated their intentions, was just 1.6 percent. Both numbers reflect the same underlying employer sentiment: 45 percent of employers characterized the overall job market for Class of 2026 graduates as “fair” — NACE’s lowest rating, and the worst employer confidence reading since 2021. The last time the largest group of employers said “fair” was in 2021, when hiring projections were also flat.
The shift from “fair” to “good” or “excellent” typically requires employers to believe that new hires will be productive quickly and that training costs are justified. The current data suggests they are skeptical. Recruiting activity has shifted later in the academic cycle — 37 percent of full-time hiring now happens in the spring, up sharply from the pre-pandemic norm where nearly three-quarters of recruiting was completed in the autumn. Employers are delaying decisions. That is not the behaviour of a market confident about what it wants.
The AI skills paradox
Demand for AI proficiency at the entry level has accelerated sharply. According to NACE’s spring update, AI skill requirements in entry-level job descriptions nearly tripled since autumn 2025. 35 percent of entry-level jobs now require AI skills, and the share of full-time postings mentioning AI has nearly doubled year over year to 4.2 percent, according to Handshake’s Class of 2026 outlook.
The Class of 2026 is, in aggregate, more AI-literate than any graduate cohort before it. Most have used AI tools in coursework, projects, and job searching. And yet NACE data shows that students remain largely unfamiliar with how to demonstrate those skills in ways that satisfy employer screening. They know how to use the tools. They do not know how to describe that use in terms an ATS or hiring manager recognises as evidence. The gap between capability and communication is where a substantial number of applications are dying.
There is also a structural irony in the numbers. AI tools are increasingly expected at entry level at exactly the moment they are compressing the number of entry-level jobs. Entry-level postings in software development and data analysis have dropped sharply, with some data indicating a 67 percent decrease in junior tech postings since 2022. The jobs requiring AI skills and the jobs being eliminated by AI are not perfectly overlapping sets, but they overlap enough to create a genuine paradox: graduates who can use the technology are competing for a reduced pool of roles partly because the technology exists.
The experience gap dwarfs everything else
The single most important variable in a 2026 graduate’s hiring prospects is not their major, their GPA, their AI literacy, or the prestige of their institution. It is whether they worked during their degree. The 81.6 percent versus 40.7 percent hiring rate differential between graduates with and without internship experience is the starkest finding in the NACE data, and it points to something employers have been trying to signal for several years without it fully landing.
Employers are not willing to pay for training at the pace they once were. The expectation, implicit in many job descriptions and explicit in others, is that a new hire can be productive quickly, without a long ramp period. That expectation is most readily met by someone who has already performed the work in an internship or co-op context. The degree signals potential. The internship signals capacity. In a market where employers are cautious, capacity wins.
For graduates already out and job hunting, this is not actionable in retrospect. For current students — and for the faculty and advisors who work with them — it is the clearest signal available about where to invest time before graduation. A semester of relevant work experience is worth more to a 2026 employer than almost any other thing a student can put on a resume.
The salary expectation gap
One final number worth stating plainly. Current undergraduates expect to earn an average of $80,004 one year after graduation. Actual starting salaries are averaging $56,153. The gap is nearly $24,000, and it is not explained by any single factor. Part of it is wishful thinking. Part of it is that salary data online skews toward high-profile industries and roles. Part of it is that students in AI and computer science fields, where starting salaries are genuinely high, make the average visible without it being representative.
The practical implication is to research salary ranges for specific roles in specific locations rather than using aggregate figures as a benchmark. A computer science graduate in San Francisco and an English graduate in Kansas City are both in the Class of 2026. Their starting salaries will have almost nothing in common, and treating the average as a target will produce frustration for most of the people doing it.