Ask a UK tech founder what keeps them up at night and the answer is rarely the product. It is hiring. Every scale-up in Manchester, Leeds or Bristol is competing for the same pool of developers, data specialists and security engineers, and that pool is not growing as fast as the demand for it. The standard response is to look upstream to universities and bootcamps. The more uncomfortable truth is that the pipeline narrows much earlier than that, in classrooms full of nine-year-olds.
The numbers behind the shortage
England rewrote its school curriculum in 2014, replacing ICT with Computing and making programming compulsory from Year 1. On paper, every child in a maintained school now learns to write and debug code before they leave primary school. In practice, the subject has been losing ground ever since.
The independent Curriculum and Assessment Review, which published its final report in November 2025, found that teaching time for Computing has declined in secondary schools since 2011/12, that there is a significant shortage of specialist Computing teachers, and that very few secondary schools teach the subject to all pupils at Key Stage 4. The review recommended replacing the current GCSE with a broader Computing qualification and exploring a new data science and AI qualification for 16 to 18 year olds.
The GCSE figures tell the same story. According to the National Centre for Computing Education, 82,428 pupils in England sat GCSE Computer Science in 2026, a 7.1 percent fall on the 88,750 entries the year before. Girls made up 23 percent of those entries. The Curriculum and Assessment Review put the participation gap more starkly still: roughly 20 percent of boys take the subject against around 6 percent of girls.
For an employer, those two figures matter more than any university statistic. A cohort that shrinks by seven percent at sixteen does not recover by twenty-two, and an industry that recruits from a population where three quarters of the candidates are male has a diversity problem that no graduate scheme can fix on its own.
Why the gap opens so early
Research into subject choice consistently shows that children form a view about whether a subject is “for them” long before they pick their GCSE options. By the time a Year 9 pupil fills in the form, the decision has often already been made, usually somewhere between the last years of primary school and the first year of secondary.
That is precisely the stage where provision is weakest. A primary teacher is expected to cover the entire curriculum, and most did not study computing themselves. Lesson time is squeezed by maths and English. Equipment is patchy. A child who happens to have a keen teacher, or a parent who works in tech, gets a head start; a child who does not may go through several years of “computing” without ever writing a line of text-based code.
The result is a lottery. Some pupils arrive at secondary school having built simple games and already comfortable with the idea that a computer does what you tell it. Others arrive believing that programming is something difficult that other people do. The second group is disproportionately female and disproportionately from lower-income households, which is exactly the pool the industry keeps saying it wants to reach.
What the technology sector already knows about learning to code
Anyone who has hired a junior developer knows that the skill is not memorising syntax. It is the habit of breaking a problem into steps, predicting what a piece of logic will do, testing it, and fixing it when it fails. Those habits can be built at eight as easily as at eighteen, and arguably more easily, because an eight-year-old has not yet decided they are bad at it.
The tools have caught up with that insight. Platforms built specifically for schools now take children from drag-and-drop blocks in Key Stage 1 through to typed CoffeeScript and Python by upper Key Stage 2 and 3, with the whole progression framed as puzzles and small games rather than lectures. CodeMonkey, an edtech platform that reports use in more than 18,000 schools worldwide, is one example: its coding curriculum for schools comes with lesson plans, automatic assessment and a teacher dashboard, which matters enormously when the person delivering the lesson is a non-specialist primary teacher with thirty other things to cover that week.
The point is not that any single product solves the problem. It is that the barrier to giving every primary pupil a structured, progressive coding experience is now cost and awareness rather than technology. That is a barrier employers are well placed to lower.
Four things a tech business can actually do
The first is the most direct. Sponsor a platform licence for the primary schools that feed your local secondaries. A school-wide subscription to a coding curriculum costs less than one recruitment agency fee, and it puts a progressive programme in front of every child in the building rather than the handful who join an after-school club.
The second is to send people, not just money. A developer spending two hours a term in a Year 5 classroom does more for a child’s sense that “people like me do this” than any careers leaflet. Programmes such as STEM Ambassadors already exist to make this easy and to handle the safeguarding checks.
The third is to be deliberate about girls. If your volunteers are all men, the visit reinforces the stereotype you are trying to break. Aim for at least half of the people you send into schools to be women, and target the transition years, Years 5 to 8, where the participation gap opens.
The fourth is to make noise. The government has accepted that the computing GCSE is too narrow and has committed to reform, but curriculum change in this country moves slowly and is easily deprioritised. Employers who write to their MP, respond to consultations and speak at regional skills boards are the reason this subject stays on the agenda.
The business case, in plain terms
None of this is philanthropy. A regional tech cluster that grows its own talent is less exposed to salary inflation, less dependent on visa policy and more resilient when the London market overheats. Companies that build visible relationships with local schools also find recruitment easier a decade later, because the graduates walking through the door have known the company name since they were ten.
The alternative is to keep competing for the same shrinking cohort and to keep paying more for it every year. The 2026 entry figures suggest that cohort will be smaller again in 2028. The children who could change that number are in primary school right now, and most of them have no idea that what they are doing with a floor robot on a Tuesday afternoon is the first step into the industry that is short of them.
Closing the gap means meeting them there.