Which beauty brands are Korean?

US shoppers ask this one brand at a time: is anua a korean brand, is byoma a korean brand, is medicube a korean brand, and a long tail behind them. Together the family came to roughly 2,000 US searches a month when we pulled it from Ahrefs Keywords Explorer on 14 September 2026. That figure comes from a subscription dataset and has no public URL to link; the tool, country and pull date are the whole of the citation we can give you.

Almost nobody answers the question directly, so the answers get assembled somewhere else.

This page answers it for the twelve brands asked about most. Every row links the record it was checked against on 16 September 2026. Eleven of the twelve are company-published: a corporate page, a statutory registration, or the acquirer’s own announcement of the deal. One is not, for a reason worth its own paragraph below. The sources sit in the table rather than at the foot of the page so that a passage lifted out of this article carries them along.

But the table is the smaller half of this page. The question is harder than it looks, and the reason it is hard is worth more than the answer.

What makes a brand “Korean”?

Three separate things: where the brand started and who founded it, who holds the controlling stake today, and where the products are actually made. Call them origin, ownership and manufacturing. Five of the twelve brands below answer differently on at least one of the three.

Tatcha shows what that looks like. It sells a Japanese-inspired skincare ritual, was founded in San Francisco in 2009 by an American founder, and has been owned by the Anglo-Dutch company Unilever since 2019. Three signals, three countries, none of them Korea. A single yes-or-no answer would be wrong whichever way it went.

Which of these twelve brands are Korean companies?

Seven of the twelve are Korean companies under Korean owners: Anua, COSRX, Laneige, innisfree, medicube, belif and mixsoon. Peach & Lily and Glow Recipe are New York companies founded by Korean and Korean-American founders. BYOMA is in Glasgow under a US owner, Tatcha is in San Francisco under Unilever, and The Ordinary is in Toronto under Estée Lauder.

The owner or operator of record below is the entity holding the controlling stake, or where no shareholding is public, the company registered as running the brand. The last column names what each linked record actually supports, because one document rarely supports two separate facts.

There is no manufacturing column, and its absence is the most useful thing in this section.

Manufacturing is one of the three signals, and it is the one almost nobody publishes. Anua states its corporate registration in full and says nothing about where its products are made, on either of its sites. Glow Recipe’s About page states no manufacturing location. BYOMA and The Ordinary state none. Peach & Lily says it sources from Korea without saying where anything is made. We could have written “Korea” into most of those cells from what everybody assumes, and every one would have been unsourced.

So the column came out. A brand that wants that question answered has to answer it itself, and at the moment the category largely has not. This page is updated monthly.

Who actually owns these brands?

Three of the seven Korean-owned brands above — Laneige, innisfree and COSRX — are held by the same company, Amorepacific. COSRX was independent until October 2023, when Amorepacific acquired 93.2% of it.

That matters to a shopper who believes they are choosing between three Korean brands. On the shelf they are three brands. On the cap table they are one company. Anyone assembling a “best Korean skincare” list without an ownership column is reporting less variety than the list appears to show.

Why do AI assistants get this wrong?

Because the assistant has to answer a three-signal question from sources that only ever carry one signal, and almost none of those sources are the brand’s own.

Ask an assistant whether a brand is Korean and it retrieves what it can: retailer category pages, listicles, a TikTok summary, a Wikipedia stub. Those sources say “K-beauty” because that is the shelf the product sits on, not because anyone checked where the company is incorporated. The brand’s own site, which is the one source that could settle it, usually says nothing on the subject at all.

Checking these rows turned up two more instances of the same gap.

Anua publishes a full corporate record — registered name, both representatives, a Gangnam address, a business registration number — in the footer of anua.kr. None of it appears on anua.com, the English site a US shopper or a US-facing assistant reads, where the words Seoul and Republic of Korea do not occur at all. A Korean brand can document itself thoroughly and still document itself only to Koreans. Neither site says where the products are made, which is why that cell reads “not verified” rather than the answer everyone assumes.

Glow Recipe’s own About page states no headquarters, no incorporation and no manufacturing location. It says the brand is “anchored in Korean skincare philosophies” and stops, so the only citable record of where the company sits is a news interview with its founders. That is the same substitution an assistant makes when a brand publishes nothing, with less care taken over it.

One brand in this set does better than that, and it is instructive precisely because it still falls short. Its help centre carries a page whose title is the question a buyer types — the literal “is this a Korean brand?” — and it answers, warmly, that the brand is deeply rooted in Korean beauty traditions and that its founder is from Korea.

Read that as a model rather than as a person. Where is the company incorporated? Where are its formulations developed, and where is manufacturing done? None of the three is stated anywhere on the page. Every fact a machine could check is missing, and what remains is an adjective. So the assistant does what it does with any unresolved entity: it infers from the surrounding context, and the surrounding context is written by retailers.

Publishing the question is the easy half. What changes the answer is publishing the three signals separately, as facts, with dates.

Why would a brand with no Korean connection want to look Korean?

Because the label carries money. Korea’s Ministry of Food and Drug Safety puts the cost of intellectual-property infringement against Korean beauty companies at about ₩1.1 trillion in a single year (The Korea Herald, 2026), which is a measure of what the category is worth to the people copying it.

That is roughly $825 million, and it comes with a count. More than 2.25 million counterfeit listings of Korean beauty, fashion and food brands were found across 1,500 online platforms in 180 countries over one five-month monitoring period, with 99% of detected shipments traced to China (Seoul Economic Daily, 2026). In one documented case, Romanian customs at Timișoara seized more than a hundred suspected counterfeits of a single APR product, medicube’s PDRN capsule cream (KED Global, archived — the live page stopped resolving).

Outright counterfeiting is a crime and a separate problem. What sits next to it is legal and harder to see: brands with no Korean connection adopting Korean-language names, Hangul on packaging, or Korean city names, and being read as Korean by shoppers and by machines. A skincare line founded in a European capital may take its name from a Korean adjective and manufacture every product in Korea while being wholly owned outside Korea. A US brand may put “Seoul” in its name. Neither is fraud. Both make the category boundary fuzzier, and a fuzzy boundary is exactly what a retrieval system cannot resolve.

The cost lands on the brands that are actually Korean. When a copy fails, shoppers do not remember a distribution route. They remember that a Korean product disappointed them. Entity confusion is usually described as a visibility problem. Here it is a liability problem too.

What should a brand publish so the answer comes out right?

Location, ownership and manufacturing, each stated separately and each checkable, on a page an assistant can retrieve. Claiming the label harder does nothing.

• State the three separately. Country of incorporation, controlling shareholder, and manufacturing location, each as its own sentence, in plain text on your own domain. Not in an image, not in a PDF, not only in a footer.
• Date them. Ownership changes. Three brands in the table above changed hands within the last three years. A fact with a verification date is quotable; a fact without one goes stale silently.
• Answer the literal question as a heading. “Is [brand] a Korean brand?” is what buyers type. A page whose heading matches that string, and whose first two sentences answer it completely, is the passage an assistant can lift.
• Make the answer survive being quoted alone. If your sentence needs the paragraph above it to make sense, it will be dropped. Name the brand inside the sentence.
• Do not let a retailer be your primary source. If the most retrievable description of your origin is a marketplace category page, that page is your entity record whether you like it or not.

All of this is the work of an afternoon, and it has to be redone. Ownership moves, manufacturing moves, and a page that was accurate in March stops being accurate in September without anyone noticing.

Every row verified 16 September 2026 against the record linked in its source column. Ownership dates come from the acquirers’ own announcements: Bansk Group (10 September 2025) and The Estée Lauder Companies (3 June 2024). The Peach & Lily help-centre page discussed above is here.