On a Sunday morning in April, a line forms outside a school gate in Hanoi. Most people in it are under 25. They carry pencils, passports and printed admission slips. Inside, they will sit the Test of Proficiency in Korean, better known as TOPIK. Last year roughly 550,000 people did the same thing in 89 countries. That number is the engine of the Korean language industry, and it has quietly become one of Korea’s oddest export businesses.
The Korean language industry is not really one market. Instead, it is four markets stacked on top of each other: a government exam, a university tuition machine, a soft-power budget line and a global app economy. Meanwhile, the people paying for all of it are mostly young, mostly Asian and mostly betting on a visa.
The version of this story that reaches foreign readers is usually the fun half. Duolingo streaks, drama subtitles, K-pop lyrics decoded on TikTok. However, the other half involves brokers on Chinese social media, students who never show up to class, and a tuition bill of 1.6 million won every ten weeks. Here is how the money actually moves.
Why 550,000 People Sat the TOPIK Exam Last Year
First, the growth curve. In 2022, about 360,000 people took TOPIK worldwide. By 2023 the figure reached 420,000. In 2024 it climbed to roughly 490,000. Then in 2025 it broke 550,000. In other words, the test more than 1.5x’d in three years.
For 2026, the National Institute for International Education scheduled 15 separate sittings. Six are paper-based, six are internet-based, and three test speaking. In addition, the online version keeps spreading to new countries, with Nepal, Laos, Bahrain and India among the recent additions. You can check the current calendar on TOPIK’s official portal, which now handles registration in dozens of markets.
Registration is not free, and the pricing tells you something. Inside Korea, the paper test costs 40,000 won for TOPIK I and 55,000 won for TOPIK II. The computer-based version runs 70,000 and 95,000 won respectively. Speaking costs another 80,000 won. As a result, even a rough estimate puts annual exam-fee revenue above 30 billion won, before a single textbook is sold.
Yet the fee is trivial next to what candidates spend preparing. That is where the real Korean language industry lives.
TOPIK, Decoded: What Six Levels Actually Buy
For readers who have never encountered it, TOPIK is graded on six levels across two papers. TOPIK I covers Levels 1 and 2. TOPIK II covers Levels 3 through 6. Crucially, you do not pick your level. Instead, you sit the paper and your score decides where you land.
Here is what each band roughly means in practice:
- Levels 1 and 2 — survival Korean. You can order food, read a subway sign and introduce yourself. For most institutional purposes, however, this band counts for very little.
- Levels 3 and 4 — the working threshold. Band 3 unlocks many undergraduate admissions and a lot of visa paperwork. Meanwhile, band 4 is what most universities want before you enter a degree program.
- Levels 5 and 6 — professional territory. These bands max out the language category in Korea’s residence point systems. In addition, they open doors in translation, teaching and corporate work.
Notably, a TOPIK certificate expires after two years. Therefore many people retake the exam not to improve, but simply to stay current. That single administrative detail generates enormous repeat demand, and it explains a good chunk of the 550,000 figure.
The Visa Engine Under the Korean Language Industry
Now for the part that outsiders consistently miss. TOPIK is not primarily a language test. Functionally, it is immigration infrastructure.
Consider the F-2-7 residence visa, awarded through a points system. Applicants are scored out of a 170-point scale and need at least 80 points to qualify. Age, education and income carry the heaviest weight, but Korean proficiency is worth up to 20 points on its own. For a mid-career applicant, jumping from Level 3 to Level 5 can be the difference between rejection and residence.
The rules also shift, which keeps the market anxious. Permanent residency, for instance, no longer accepts TOPIK at all. Since March 2019, F-5 applicants have had to clear the government’s Korea Immigration and Integration Program instead. Consequently, a certificate that once bought long-term security now buys something narrower.
Employment adds another layer. Skilled-worker categories, regional visas and the manufacturing talent pipeline all attach language conditions of some kind. Korea’s broader work visa overhaul, which compresses 39 categories into three, keeps proficiency at the center of the scoring. Even the labor reform push for foreign talent treats Korean ability as a proxy for how likely someone is to stay.
So the incentive structure is clean. Learn Korean, gain points. Gain points, gain time. Gain time, gain a life here. Above all, that is why demand for the TOPIK exam tracks immigration policy far more closely than it tracks K-pop.
The Other Korean Exam Almost Nobody Writes About
There is a second Korean test, and in human terms it may matter more.
EPS-TOPIK is run by HRD Korea for the Employment Permit System, the channel that brings in E-9 workers for factories, farms, construction sites and fishing boats. Seventeen countries send workers through it. Crucially, it has nothing to do with the exam that university students take. A pass earns no F-2-7 points and no academic credit whatsoever.
Instead, it does something blunter. Candidates are ranked by score, and the roster is cut off at the national quota for each sector. Consequently, EPS-TOPIK is not really a proficiency test at all. Rather, it is a lottery with vocabulary attached.
The economics are brutal in the best sense. Registration costs roughly $28, which is a fraction of a regular TOPIK sitting. In return, a passing candidate can access wages several times higher than at home. Therefore entire preparation industries have grown up in Kathmandu, Jakarta and Phnom Penh, complete with dormitory cram schools, mock-test apps and Nepali-language textbooks that outsell most Korean novels.
Scale gives you the picture. Nepal alone opened a 2025 EPS round for about 5,300 worker slots, and applications ran far beyond that. Similar rounds repeat across all seventeen countries every year. So while the academic Korean language industry gets the headlines, the labor track quietly processes a comparable flow of people. Anyone curious can browse the official EPS-TOPIK portal, which publishes schedules and past papers in multiple languages.
Two tests, then, and two very different Koreas at the end of them. One leads to a Yonsei classroom and a possible residence visa. The other leads to a five-year cap on a factory floor. Both, however, begin with the same alphabet.
Vietnam Overtakes China: A New Map for Korean Language Education
In 2026, Korea passed a milestone it had been chasing for years. Foreign enrollment at Korean higher education institutions hit 307,464 students, up 54,030 in twelve months. That is a 21.3% jump in a single year.
The government’s Study Korea 300K plan had targeted 300,000 students by 2027. Korea got there in 2026 instead, a full year early. Nevertheless, the composition of that number matters more than the number itself.
For the first time ever, Vietnam displaced China at the top:
| Rank | Country | Students | Share | Year-on-year |
|---|---|---|---|---|
| 1 | Vietnam | 96,834 | 31.5% | +28.9% |
| 2 | China | 78,890 | 25.7% | +3.1% |
| 3 | Uzbekistan | 20,876 | 6.8% | +32.2% |
| 4 | Nepal | 20,118 | 6.5% | +57.4% |
| 5 | Mongolia | 17,189 | 5.6% | — |
Vietnam alone accounted for 40.1% of the total increase. Meanwhile, Chinese enrollment grew barely 3%. Nepal’s 57% surge is arguably the more striking signal, because it points to a second wave of source countries coming online.
At the same time, concentration is easing. The top three nationalities made up 66.1% of foreign students a year ago. Now they account for 64.0%. In short, Korean language education is diversifying its customer base faster than most people realize.
Immigration data fills in the rest of the picture. As of mid-2025, Korea hosted 272,573 residents on student visas. Of those, 190,061 held the D-2 degree visa, while 82,396 held D-4-1 language trainee status. That second group is the pure fuel of the university language institutes.
Sejong Institutes: The Korean Language Industry as Soft Power
Not everyone learning Korean comes to Korea. Many start at a King Sejong Institute, the government-backed network that functions like Korea’s Alliance Française or Goethe-Institut.
The scale-up has been dramatic. In 2007 the program launched with 13 centers and 740 students. Today it operates 252 institutes across 87 countries, serving more than 210,000 learners online and offline. When 11 new institutes opened in nine countries in 2025, the Korea Times reported an application ratio of 8.5 cities competing for each slot.
That ratio is the interesting data point. Local institutions are lobbying to host Korean classes, and Seoul gets to choose. Few countries enjoy that position with any language.
The next phase expands regional hub institutes from 5 to 20. Those hubs train teachers, localize curriculum and manage nearby centers. Therefore the network is shifting from broadcasting to franchising, which is a far cheaper way to grow.
Still, soft power has a budget, and budgets have politics. Every won spent on Korean language education abroad is a won not spent at home, where the private education market already absorbs about $20 billion a year. So far the return has been easy to defend. Whether that holds through a leaner fiscal cycle is a genuinely open question.
Who Makes Money in the Korean Language Industry
Follow the cash and four groups appear.
University language institutes. These are the quiet giants. Yonsei’s Korean Language Program, for example, charges 1,600,000 won per ten-week level, plus an 80,000 won application fee. Each level runs 200 hours, and finishing all six takes roughly a year and a half. Multiply conservatively: 82,000 language-visa students, three terms a year, 1.4 million won per term. That alone approaches 350 billion won, or about $250 million annually.
Universities themselves. Regional campuses facing a domestic demographic collapse have found foreign tuition irresistible. Korea’s birth rate has hollowed out freshman classes, so international students increasingly keep smaller schools solvent. Consequently, recruitment has become a survival strategy rather than an internationalization strategy.
Agencies and recruiters. Between a village in Nepal and a classroom in Daegu sits a chain of intermediaries. Some are legitimate. Others charge thousands of dollars for services that a free Study in Korea portal already provides.
Apps, publishers and creators. This tier is the fastest-growing and the least measured. Global Market Insights valued the Korean language learning market at $7.2 billion in 2024, with an aggressive forecast toward $67 billion by 2034. Treat the long-range number with caution. However, the direction is not really in dispute, and learners aged 18 to 24 already represent more than half of the demand.
The KPop Demon Hunters Effect on Learning Korean
Culture does not create the Korean language industry. Instead, it fills the top of the funnel, and 2025 delivered an unusually large fill.
Duolingo’s 2025 Language Report put Korean sixth among the world’s most-studied languages, overtaking Italian. Part of that came from access rather than fashion: for the first time, speakers of more than 20 languages could study Korean from their own native language instead of routing through English, Japanese or Chinese. In addition, Korean ranked as the second fastest-growing language in seven countries, including Argentina, France, Germany, Mexico and Poland.
Then came the animated film that ate the internet. KPop Demon Hunters pulled five Grammy nominations at the 68th Grammy Awards, and “Golden” won Best Song Written for Visual Media. Afterward, the Korea Times reported that Duolingo’s American Korean learners jumped 22% in a year.
The classroom data is even more telling. Korean enrollment in US higher education rose 38% between 2016 and 2021, while overall foreign language enrollment fell 16%. As Columbia University’s Korean program director put it, Korean is simply “cool” now.
Of course, cool is not the same as committed. Most people who download an app after watching a music video never reach Level 3. Nevertheless, a small conversion rate applied to a very large audience still produces a meaningful pipeline. Fandom economics work the same way in K-pop’s trainee system: millions at the entrance, a few thousand at the exit.
The Dark Side of the Korean Language Industry
Any test that controls immigration eventually attracts fraud. TOPIK is no exception, and the scale is uncomfortable.
A Korea Herald investigation documented Chinese brokers openly advertising on Xiaohongshu. Their menu included micro-earpieces for real-time assistance at around 20,000 yuan, roughly 4.3 million won. Advance answer keys ran 25,000 to 35,000 yuan, or up to about 7.5 million won depending on how early you wanted them. Sellers promised 90% accuracy and refunds if a proxy test-taker failed.
One method was almost elegant. Candidates sitting the exam in earlier European time zones memorized questions, then sold them onward to brokers who resold to Asian candidates hours later. At least one student confirmed that posted answer keys matched the real paper.
Officials at the National Institute for International Education acknowledged the problem. However, they described prevention as “realistically difficult,” citing rising candidate volumes and the widening range of institutions that accept TOPIK scores. That is a candid answer, and also an alarming one.
Classroom quality is the second problem. At Korean junior colleges, fewer than 30% of international students meet the expected Korean proficiency standard. One university president publicly described schools that hold a single day of in-person class, shift everything else online, and then leave students effectively unattended. In those cases, tuition arrives and education does not.
Overstays complete the picture, though the numbers are less dramatic than headlines suggest. Among D-2 degree students, new illegal stayers rose modestly to 1,196. Among general D-4 trainees, the figure actually fell to 1,302. So the system is leaking, but it is not hemorrhaging. The deeper risk is reputational: if employers stop trusting Level 4, the whole credential loses value.
Where the Korean Language Industry Goes Next
Three forces will shape the next five years.
First, demographics. Korea needs workers, and language is the cheapest filter available for choosing which ones to admit. Therefore proficiency requirements will keep tightening rather than loosening, and the exam will keep growing alongside them.
Second, verification. Once brokers can sell answers, the paper test becomes a liability. Expect faster migration toward internet-based and speaking assessments, since both are much harder to game. The 2026 schedule, with six IBT sittings and three speaking sittings, already points that way.
Third, competition for students. Japan, Taiwan and Germany are recruiting from the same countries with similar promises. Korea’s advantage is cultural pull, and its disadvantage is a reputation for treating foreign students as revenue. Programs like the digital nomad visa suggest the government understands it must sell a life, not just a classroom seat.
Meanwhile, there is a domestic irony worth noting. Korea spends enormously on English education and stages a national shutdown for the CSAT every November. Now the same test-obsessed machinery has been pointed outward, and 550,000 foreigners are sitting Korea’s exam instead.
If You Are Learning Korean in 2026: A Practical Route Map
Finally, some practical guidance, since many readers are inside this market rather than observing it.
If you want a degree. Target Level 3 for admission and Level 4 before your first semester. Budget a year at a university language institute, and expect roughly 5 to 7 million won in tuition plus living costs. Apply through the official Study in Korea portal before paying any agency.
If you want to work. Level 4 is the practical floor for most office roles. Level 5 or 6 changes your negotiating position entirely, especially in translation, teaching and customer-facing work. Remember the two-year expiry and time your sitting accordingly.
If you want to stay long-term. Track the points table rather than the exam alone. For F-2-7, language is capped at 20 points out of the 80 you need. For permanent residency, the immigration integration program has replaced TOPIK, so plan for that pathway separately.
If you are learning for fun. Ignore all of the above. Watch the show, learn the lyrics, and take the exam only if a certificate genuinely helps you. Roughly half the world’s Korean learners never sit TOPIK at all, and they are not doing it wrong.
The Bottom Line
The Korean language industry looks like culture from the outside and behaves like immigration policy from the inside. Its growth is real, its revenue is concentrated in universities and agencies, and its weakest point is trust. Fix the fraud and the neglected classrooms, and Korea has a durable global asset. Ignore them, and a 550,000-person exam slowly turns into a 550,000-person paperwork ritual.
For now, the line outside that school in Hanoi keeps getting longer.
