Korea's Top Science Schools Top 30,000 Applicants on AI Hype

What happened

Applications to South Korea's four national institutes of science and technology — KAIST, GIST, DGIST and UNIST — passed 30,000 for the first time in the early-admissions (susi) round for the 2027 academic year, according to tallies from the private education analytics firm Jongro Hagwon cited in Korean press reports. It is the highest figure on record.

Over the same cycle, combined early-admission seats at the four schools rose from 1,727 to 2,067 — an increase of 340. Most of that expansion comes from newly created AI departments, alongside rising preference for semiconductor-focused majors.

The timing is what makes this notable. Korea is simultaneously expanding medical school enrollment through a new regional-physician quota system. Yet instead of draining top-scoring students away from engineering, this cycle saw science-track applications hit a record. Reports attribute the shift to industrial outlook for AI and chips, plus well-publicized bonus payouts at major semiconductor firms — Korean media regularly covers chipmaker performance bonuses running into hundreds of thousands of dollars.

Why now

For roughly a decade, the default choice for Korea's highest-scoring high school students has been medical school. That held even as universities rolled out semiconductor majors backed by corporate hiring pipelines. This year's surge is the first clear quantitative hint that the pattern may be loosening.

Two forces overlap. On the supply side, the institutes actually created capacity: 340 additional seats means there is somewhere to apply. Preference alone doesn't move application counts without seats. On the demand side, constant news flow about AI infrastructure spending and a recovering memory-chip market has raised what students and parents expect to earn after graduation.

How this differs from the semiconductor-major wave

The earlier expansion of chip-focused programs ran mostly through corporate contract departments, which sold guaranteed employment at a specific company. The new AI departments are structurally different: they add undergraduate capacity at research-oriented institutes, not a narrow pipeline into one employer. That points toward a researcher track rather than pure industrial workforce supply.

There is also a mechanical caveat. Unlike regular Korean universities, the four institutes fall outside the six-application cap that governs early admissions. Students can apply to them in addition to six ordinary universities. So the 30,000 figure overstates true preference — many applications are effectively insurance. The real signal will be where dual-admits actually enroll.

What actually changes

For industry, the relevant horizon is not this year's hiring market but the talent pipeline four to six years out. Today's AI freshmen reach the labor market in the early 2030s. Whether the current investment cycle persists — or behaves like semiconductors, with sharp boom-bust swings — will determine how crowded that cohort feels. Korea has repeated this pattern before, expanding majors during shipbuilding and construction booms that later cooled.

For applicants, two practical points stand out: first-year AI departments have no admissions history, so cutoff lines are hard to predict; and because institute applications sit outside the six-school cap, they can be used strategically at low cost. For policymakers, the collision of expanded medical quotas and expanded AI quotas in the same year means dual-admit enrollment data will become the key evidence in Korea's long-running argument over where its top students should go.