AI Makes This the Best Time to Study Maths. Degrees Are Getting Harder to Reach.

Maths is once again the most popular A-level subject. There were 115,380 entries in 2026, an increase of 3.9% on 2025, and further maths AS-level entries rose 8.5%.

Ucas data for the 2026 application cycle showed an 11% increase in applications to study mathematical sciences at university.

Students are choosing the subject in greater numbers, and choosing to continue with it.

Why now is a good moment

The case for studying mathematics has strengthened rather than weakened as generative AI has improved, which is the opposite of what most people assume.

AI systems are solving mathematics problems that have stood open for years. That is a genuine and startling development. It is also a development that requires mathematicians — to formulate the problems, to verify the proofs, to judge which results matter, and to understand well enough what the systems are doing to know when they are wrong.

A machine that produces a proof does not produce the judgement required to evaluate it. That judgement is the discipline.

The contraction

Against rising demand, university mathematics provision is going the other way.

Departments have closed or merged. Course offerings have narrowed. The financial model of English higher education — a frozen fee cap against a decade of inflation — has made it difficult to sustain subjects that are expensive to teach well and do not attract large international cohorts.

Mathematics is comparatively cheap in equipment and expensive in staff, because small-group teaching is how the subject is actually learned. That is exactly the cost structure that suffers when institutions economise.

What less accessible means in practice

Accessibility is not principally about entry grades, though those are rising.

It is about geography: a student who cannot relocate has fewer local options each year. It is about school provision, since further maths remains unavailable at many state schools, and its absence disadvantages applicants at selective universities. And it is about cost, in a system where three years of living expenses is a larger barrier than tuition.

Each of these filters the intake in the same direction, and none of them measures mathematical ability.

The mismatch that should worry people

There is an unusual alignment here that policy is failing to exploit.

A country wanting to build capability in AI, cryptography, modelling, finance and data infrastructure needs mathematicians, and demand for them is rising in both the labour market and the classroom. School students are volunteering in record numbers. The bottleneck is entirely in the middle: the capacity of universities to teach them.

That is the rare kind of problem where the constraint is identified, the demand is proven, and the remedy is funding a subject that is not especially expensive to provide. It is being handled instead by allowing departments to close and letting the applications compete for a shrinking number of places.

What is lost when a department closes

A mathematics department is not a set of modules that can be reassembled later.

It is a group of specialists whose research areas overlap enough to supervise students across a full curriculum, assembled over decades. When one closes, the staff disperse into industry or abroad, and the region loses not only the degree but the school outreach, the teacher training pipeline and the local supply of graduates who might have taught the next cohort.

That is why the closures matter more than the headline numbers suggest. Places can be added back to a surviving department within a year. Rebuilding a department that has gone takes considerably longer than the shortage it was closed during.