Actually Japan What student life here is really like

HomeActually JapanBy country

By country

Nepal has more than doubled, and vocational schools have overtaken universities — both in two years

In May 2023 there were 279,274 international students in Japan. In May 2025 there were 408,069. That is a big number, but the size is the least interesting part: in the same two years the largest single group changed, and so did the most common type of school. Most guides to studying in Japan are describing the map from before.

Japan counts its international students once a year, on 1 May. The two most recent counts are two years apart in our reading here, and putting them side by side shows something that a single year's figures hide.

Change in students by country, May 2023 → May 2025
Myanmar×3.78
Nepal×2.65
Vietnam×1.19
China×1.14
South Korea×0.98

Multiplier over two years. The national total grew ×1.46 in the same period. Anything below that line is losing share even while growing.

In absolute numbers: Nepal went from 37,878 students to 100,239, an increase of 62,361 people. China, still the largest group by a distance, added 15,604. Nepal added four times as many students as China did, from a base one third the size.

CountryMay 2023May 2025Change
China115,493131,097+13.5%
Nepal37,878100,239+164.6%
Vietnam36,33943,366+19.3%
South Korea14,94614,612−2.2%
Myanmar7,77329,413+278.4%
All countries279,274408,069+46.1%

Nepal and China together are now 56.7% of every international student in Japan. Two nationalities, more than half the population.

South Korea has fewer students in Japan than it did two years ago. In a period when the total grew by nearly half, that is not stagnation — it is a different country's story entirely.

The second change is bigger and nobody talks about it

The same count breaks students down by what kind of institution they attend. Here the movement is not in the countries at all:

Type of institutionMay 2023May 2025Change
Vocational school (senshu gakko)46,325106,829+130.6%
Junior college1,9554,138+111.7%
Japanese language school90,719140,174+54.5%
University (undergraduate)80,36292,442+15.0%
Graduate school55,53960,013+8.1%
Technical college501527+5.2%
Preparatory course3,8733,946+1.9%

Read the first and fourth rows together.

May 2023

+34,037

Undergraduate degrees had that many more international students than vocational schools.

May 2025

+14,387

Vocational schools now have that many more than undergraduate degrees. The order reversed.

The most common destination for an international student in Japan, after the language school, is no longer the university. It is the vocational school — 106,829 students against 92,442. Two years ago the gap ran the other way by more than 34,000.

Growth of 130.6% in two years is not a trend, it is a redirection. We are not going to explain it, because the count does not come with an explanation and any reason we offered would be a guess. But if you are choosing between routes, it is worth knowing that the route you are being sold is the one that just doubled.

Where each country goes

The two questions combine. JASSO publishes the country breakdown separately for language schools and for higher education, and the split by nationality is not remotely uniform:

Share of each country's students who are at a Japanese language school, May 2025
Myanmar55.0%
Sri Lanka45.9%
Vietnam45.4%
Nepal44.5%
Bangladesh36.4%
Taiwan25.7%
China24.7%
Indonesia19.1%
South Korea7.4%

Our subtraction: total students minus those at higher education institutions. National figure 34.4%.

Nearly half of everyone arriving from Nepal is at a language school. Fewer than one in thirteen arriving from South Korea is. Those two groups are having such different experiences of Japan that most general advice about "international students" applies cleanly to neither.

What the language school year actually involves, we have written about separately: 49.7% of those students work twenty hours a week or more, 13.2% receive a scholarship, and the average scholarship is ¥35,000 a month. Against the undergraduate figures — 20.4% and 51.5% — it is a materially different year.

Now the uncomfortable part, about our own sources

Almost everything else on this site comes from JASSO's living-conditions survey: what students spend, how they work, whether they were asked for a guarantor. It is the best data of its kind in Japan and we rely on it heavily.

It was conducted between November 2023 and March 2024.

What that means for every number we publish

The living-conditions survey describes a Japan with roughly 279,000 international students. The current figure is 408,069. The population it measured has since grown by nearly half, and the fastest-growing parts of it are exactly the parts that survey covers least — the vocational schools, and students from Nepal and Myanmar.

How far off does that make it? We can put a bound on part of it. In the survey's own sample, Nepal is 10.1% of respondents; in the actual May 2023 population it was 13.6%. China is 38.1% of the sample against 41.4% in the population. Those gaps are modest, and some of it is explained by the survey deliberately excluding government-funded and short-term exchange students. So the survey is not badly drawn for the year it was taken.

The problem is the two years since. We have no way to correct for that, and neither does anyone else quoting the same figures. When we cite spending or working hours, read them as 2023 conditions for a smaller and differently-composed student body — reliable about the structure of student life, less reliable as a snapshot of today.

The counts on this page are from May 2025 and do not have that problem.

What to take from this

Sources. Japan Student Services Organization (JASSO), 2025 Survey of International Students in Japan (as of 1 May 2025) — result page, full data PDF. And the 2023 Survey of International Students in Japan (as of 1 May 2023) — result page. Living-conditions figures and the sample composition: FY2023 Survey on Living Conditions of Privately Financed International Students, Table 1-2 (respondents by country) and the survey method section, which gives the fieldwork dates and the exclusion of government-funded and exchange students — PDF. Our calculations: all two-year multipliers and percentage changes are ours, computed from the two published counts. The language-school shares by country are total minus higher education; the two published totals sum to exactly 408,069, which is what makes that subtraction safe. The +34,037 and +14,387 comparisons are simple differences between the vocational and undergraduate rows.

One finding a week, from the data

No advice, no agency, nothing to sell you. Just what the numbers say about living here.

What we do with your address — and the list of what we do not collect.