Class of 2026 facing hiring squeeze
Stanford’s student paper reports that the class of 2026 is finding it unusually hard to land entry‑level computer‑science jobs, describing a reversal from a few years ago. The piece argues this is not solely AI‑driven but reflects broader recruiting caution, which raises the premium on clear, demonstrable skills and independent project evidence. (stanfordreview.org)
Stanford students who expected the usual conveyor belt from computer-science degree to software job are finding that the belt has slowed to a crawl in 2026, with the Stanford Review calling this one of the toughest recruiting seasons in recent memory and citing Stanford professor Jan Liphardt saying entry-level hiring has reversed sharply from three years ago. (stanfordreview.org) This is not just one campus having a bad semester. The National Association of Colleges and Employers said employers plan to raise overall hiring for the class of 2026 by only 1.6%, and 45% of employers rate the market for new graduates as merely “fair.” (naceweb.org) The squeeze is even tighter in software, where the old pattern was to hire juniors in bulk and train them on the job. Indeed data published through the Federal Reserve Bank of St. Louis shows U.S. software-development postings at 72.44 on a February 2020 equals 100 index as of March 27, 2026, which means postings were still roughly 28% below their pre-pandemic baseline. (fred.stlouisfed.org) Indeed’s Hiring Lab says the tech hiring freeze had already entered its third year by mid-2025, and tech postings on Indeed were 36% below their February 2020 level in July 2025. The same analysis says nearly half of the net decline from the peak happened before ChatGPT 3 went public, which undercuts the idea that one chatbot suddenly erased the whole entry-level market. (hiringlab.org) SignalFire’s 2025 talent report describes the same shift inside company rosters: new graduates accounted for just 7% of hires at large tech companies and under 6% at startups, with new-grad hiring down more than 50% from 2019 at big firms. That is less like a door closing because robots arrived and more like companies deciding they would rather hire fewer people who can contribute on day one. (signalfire.com) Recruiters are also moving more decisions later because they are less sure about budgets. The National Association of Colleges and Employers says 37% of full-time college recruiting for the class of 2026 is happening in the spring, compared with the old pre-pandemic pattern when nearly three-quarters of full-time recruiting was done in the fall. (naceweb.org) That change hits computer-science students especially hard because the old campus playbook depended on predictable fall recruiting, internship pipelines, and brand-name new-grad programs. When those programs shrink or slide into spring, students with no prior internships or shipped products can spend months applying into a market that has fewer openings and slower answers. (naceweb.org) The Stanford Review argues the winners in this market are the students who can point to something concrete: a working app, a research project, open-source code, or an internship where they solved a real problem. The National Association of Colleges and Employers says nearly 70% of employers now use skills-based hiring and that employers want candidates to share specific examples of solving problems, while nearly all respondents said U.S.-based internships are valuable. (stanfordreview.org) (naceweb.org) Artificial intelligence is still part of the story, but mostly as an amplifier. Indeed says jobs tied directly to artificial intelligence, such as machine-learning engineer roles, are among the few tech jobs still above early-2020 levels, which means demand has not vanished so much as shifted upward and narrowed toward people with rarer skills. (hiringlab.org) That is why students now hear a harsher version of an old hiring rule: “show me” beats “I studied it.” In a market with flat graduate hiring, fewer junior software postings, and more skills screening, a transcript opens fewer doors than a portfolio that proves you can already build something useful. (naceweb.org) (fred.stlouisfed.org)