Leadership1 publisher3 min readPublished
Coding placements for UK computer science graduates fell to 28% last year
The Guardian's 2027 university guide puts 28% of computer science graduates in professional coding jobs last year, down from about 40%, and the two analysts reading the data disagree over how much of the drop AI explains.
The Board Room · Leadership desk
What happened
- Matt Hiely-Rayner, who compiles the Guardian University Guide, says 28% of UK computer science graduates went into professional coding or programming jobs last year, down from about 40% previously.
- The share of computer science graduates entering any graduate-level occupation fell from more than 60% two years ago to 50%.
- Coding and software development were the fastest-falling graduate occupations in the data compiled for the 2027 edition of the guide, published on Saturday.
- The figures come from the Higher Education Statistics Agency graduate outcomes survey, which asked more than 350,000 former students about their careers 15 months after they finished courses in 2024.
Compiled by The Board RoomSomething wrong?How this is made
Why it matters
- decision Anyone sizing a 2026 graduate intake has to choose while the cause of last year's fall is contested, and the next cohort's outcome data will not arrive before the offers go out.
- contradiction Hiely-Rayner reads the fall as AI supplying cheap grunt work; Ball says there is little evidence AI did it. A headcount plan justified by AI substitution is resting on the contested half of the record.
- constraint Course redesign can add AI content, but the placement rates that judge it are measured 15 months after graduation, so a programme changed this year first shows up in outcome data several years out.
- exposure Applicants picking a computer science course this autumn are choosing on outcomes recorded for people who finished in 2024, two admissions cycles before their own.
Subtract the coding share from the overall share and the fall looks specific to one job. Graduate-level work outside coding accounted for about 22 points of last year's 50%, against about 20 points of the 60%-plus recorded two years earlier [16]. The coding placement rate fell by 12 points. The decline in graduate-level employment of any kind was at least 10 points [17]. Against its own base, the coding rate lost about 30% [18]. The guide's two baselines are not dated to the same year, so the subtraction is indicative.
The two analysts quoted on the figures do not agree on cause. Matt Hiely-Rayner, who compiles the guide for the consultancy Intelligent Metrix, said: "With AI providing such responsive and cheap grunt work in this field, it's hard not to conclude that this is behind the trend" [4]. Charlie Ball, head of labour market intelligence for Jisc, said: "It is clear that something did happen to the software developer labour market last year. While I'm reluctant to say that AI did it, because there's little evidence, it's very hard to escape the conclusion that AI has played a role" [6].
The objection to reading this as displacement is a fair one: a single cohort in a single country, surveyed once at 15 months out [5], with the AI attribution unproven on Ball's own account [6]. What the record supports is narrower than displacement. Ball said the data "shows the signs of an industry that is changing, and graduates are feeling the sharp end of that because they are the largest group of new employees" [8]. He also said: "It's not collapsed entirely, there are still roles available, but pure coding jobs seem to have been in particularly short supply in 2025" [9].
In finance, Ball describes employers changing jobs instead of cutting them. He said: "I've talked to a lot of finance employers about their use of AI, and most saying they are not replacing jobs with it, they are changing jobs. They expect new hires to have familiarity with AI" [10]. He added that the field is fast-moving and employers are not unanimous: "Some have an appetite for risk and are embracing AI tools and laying people off, in the hope that it will give them an edge. But that may not persist, because everyone is experimenting right now" [11]. Employer demand for graduates in well-paid roles such as economists and management consultants also declined [12].
On the graduate side the visible adjustment is lateral. Ball said there are signs computer science graduates are diversifying into related roles such as cybersecurity and network engineering [7]. Universities are moving on course design: the University of Birmingham, which rose to 14th overall in the new tables on improved student satisfaction [13], now lets undergraduates add an intercalated year in subjects such as AI and data science without the prerequisites normally needed [15]. Its vice-chancellor, Adam Tickell, said: "This is about equipping people for the world they are in, whether or not we like the world we are becoming" [14].
For an employer the tradeoff is between two dated costs. Taking juniors now means paying for eighteen months of work an AI tool can partly supply, on the expectation of a mid-level engineer in 2029. Skipping the intake means buying that engineer later from a cohort in which 28% went into coding jobs at all [2]. The first cost sits on this year's payroll; the second sits on 2029's.
What to watch
- The next graduate outcomes survey, covering 2025 completers, will show whether 28% was one cohort or a new floor.
- Whether cybersecurity and network engineering placements absorb the diverted graduates in the next round of subject-level data.
- Whether the finance employers Ball describes laying people off keep the reduced headcount once the experimentation phase ends.