How Platform Power Loses Its Own Legitimacy

The End of the Naver Era

- How Platform Power Loses Its Own Legitimacy

Once, opening the internet in Korea meant opening Naver. Searching, reading the news, reading blogs, entering cafés, shopping, looking at maps, finding people and companies, buying something — almost every one of these took place inside a single platform. Naver was not merely a website. It was the interface that organized the information life of Koreans, and at the same time the economic gateway that advertisers, sellers, and content producers had to pass through in order to reach users.

That era has already begun to end.

What matters is not whether Naver collapses right away. It is clear that Naver still makes money: in the second quarter of 2026 it recorded revenue of 3,388.8 billion won and operating profit of 520.3 billion won. Revenue grew 16.2% year on year. So calling Naver a financially collapsed company today is simply not true.

But a company's historical standing is not decided by its income statement alone.

The more important question is this.

Do Koreans still absolutely need Naver?

The answer to that question is changing.

In July 2026, the Google app's monthly active users in Korea reached 47.02 million, surpassing Naver's 46.85 million for the first time. Of course, this figure is not the same as search share, and it excludes PC web usage, so it alone does not mean Naver search has stepped down from first place. But at the very least it clearly shows one change: the internet touchpoints of Koreans are no longer concentrated on Naver alone.

The more important change is not even the simple competition with Google.

Search itself is passing over to AI.

In the past, a user typed in a query, read the results, and judged for themselves. Now the user can ask a question, explain the context, and have the AI synthesize many sources. What the user wants is no longer “search results” but “the problem solved.”

For Naver, this change is a threat that goes beyond a mere swap of competitors.

It is not that Google replaces Naver;

it is that the search interface itself may be replaced by an AI interface.

1. Naver's problem does not end with quality

There are two kinds of reasons a user comes to dislike a service.

One is simply that the product is inconvenient.

The other is that there are problems both in the process of using the service and in the results the service produces.

To understand the criticism of Naver properly, you have to look at the latter.

Naver's problem does not end with the personal taste of “I don't like the search results.”

Looking at the incidents actually confirmed around Naver over the past several years, at least four power relations exist.

Organizational power over workers.

Platform power over businesses and competitors.

Informational power over consumers' choices.

Interface power through search and recommendation algorithms.

And these four are not independent of one another.

A giant platform controls information through technology, gathers users through information, draws in businesses through users, and through that network expands its economic and organizational power once again.

So the most important question in evaluating Naver is not “is it a good company or a bad company?”

It is how a company holding that much power has exercised it.

2. Labor rights: the 2021 special labor inspection was the decisive event

The clearest evidence is the special labor inspection by the Ministry of Employment and Labor.

After an employee at Naver died in May 2021 while suffering from workplace harassment, the Ministry of Employment and Labor conducted a special labor inspection of Naver.

As a result of the inspection, the ministry concluded that the deceased worker had endured continuous verbal abuse and humiliating remarks from a direct superior, had been deliberately excluded from decision-making, and had been under excessive work pressure. Naver's management was found to have received complaints from multiple employees without properly investigating the facts.

More serious was how the company handled things after the report.

Far from protecting the worker who reported the workplace harassment, the ministry's inspection confirmed a case in which the company assigned the victimized worker to a temporary department unrelated to their existing work and gave them no duties. The ministry judged this to be adverse treatment.

This is not simply a matter of “a boss said bad things.”

The most important protective system in an organization looks like this.

Problem occurs → Report → Investigation → Protection → Correction

But at Naver at the time, this chain did not work properly.

A problem occurred,

a report was made,

the company was aware of the fact,

and yet the government investigation confirmed that no appropriate investigation or protection took place.

3. The organizational-culture problem was also hard to see as one individual's case

During the special inspection, the Ministry of Employment and Labor also conducted an organizational-culture survey.

Of 1,982 respondents, 52.7% said they had experienced workplace harassment at least once in the previous six months.

10.5% said they experienced it repeatedly, at least once a week.

There is an even more shocking number.

Among those who had experienced harassment, 44.1% said they mostly endured it alone, and only 6.9% said they appealed to the company or a superior. The most common reason was “because it didn't seem like it would be resolved even if I responded.”

Of course, this survey does not mean that all 52.7% were legally recognized victims of workplace harassment. It was a survey of perception and experience.

But precisely for that reason, it is data that shows the state of the organizational culture.

More than half of respondents felt they had experienced harassment, and a large share of them believed that raising the issue would not resolve it.

That is a strong signal that trust in the normal protective system had been substantially damaged.

4. The labor-law violations were a separate problem

Harassment was not the only thing the special inspection found.

The Ministry of Employment and Labor found that over the previous three years Naver had failed to properly pay overtime, night, and holiday allowances to current and former employees. The unpaid amount announced at the time was about 8.67 billion won. It was also confirmed that 12 pregnant workers had been made to work overtime.

Naver's subsequent disclosures record that, following the special inspection results:

- A fine of 19.36 million won was paid

- 8.516 billion won in additional overtime allowances and the like was paid

- 16 dispatched workers were converted to direct employment

That is what the disclosures record.

What matters here is not only the size of the amounts.

In a single special inspection, distinct labor problems — workplace harassment, the reporting and protection system, wages and working hours, protection of pregnant workers, and dispatched labor — all surfaced at once.

So it is hard to reduce the 2021 incident to merely “an unfortunate individual case.”

This is why it becomes possible to conclude that the organization's operations and its labor-management system itself were the problem.

5. But didn't Naver improve?

It did.

This point must also be acknowledged.

After the incident, Naver overhauled its procedures for reporting and remedying workplace harassment and pursued reforms such as an investigation and review structure involving labor, management, and outside experts.

Naver also stated at the time that it “feels a heavy responsibility” for the special inspection results and said it would prepare a comprehensive transformation. At the same time, it said some of the findings differed from the facts and that it would offer explanations.

So it is not accurate to say that the Naver of 2026 is exactly the same organization as in 2021.

But it does not end here.

In 2025, the trust problem erupted again.

6. The 2025 return of Choi In-hyuk showed the “memory of the organization”

In 2025, Naver brought former COO Choi In-hyuk back as head of its Tech Business division.

The former COO had been COO in 2021, when an employee died from workplace harassment at Naver, and had afterward stepped down as a member of management, taking moral responsibility.

Naver's union pushed back hard.

The union-wide vote in May 2025 covered 5,701 members, and 4,507 participated.

Of those, 4,454 — 98.82% — opposed the former COO's return.

Turnout was 79.06%.

This number should not be used as if it were a legal judgment.

Nor does it mean that 98.82% judged that “Naver is currently violating labor law.”

But what it shows is clear.

The 2021 incident and the question of responsibility still remained inside the organization as a problem of trust.

Organizational culture does not change with a rulebook alone.

The members of an organization have to be able to believe in the conduct of management.

In that sense, the 98.82% of 2025 is not a number that proves Naver's legal liability. It is a strong signal of how badly organizational trust had been damaged.

7. Beyond the labor problem, Naver's second power appears

Market power.

Naver is not a simple content company.

It gathers sellers,

provides search,

exposes products,

sells advertising,

and connects transactions.

In this structure, the platform is not a mere intermediary.

It becomes a gatekeeper controlling the entrance to the market.

And there are actual cases.

8. The Naver real-estate case: a platform that controlled the movement of information

The Korea Fair Trade Commission took issue with Naver's conduct in contracting with real-estate information providers on terms that prevented them from supplying listing information to the rival platform Kakao.

The commission imposed a corrective order and a penalty surcharge of 1.032 billion won.

The essence of this case is not a simple contract term.

It is that the platform decided where information could flow.

When a platform sits between businesses that have information and consumers who want it, the platform can control the flow of that information.

And by keeping information from reaching a rival platform, it can degrade the very quality of the rival's service.

That is platform power.

9. The shopping-algorithm case was far more fundamental

Naver Shopping showed a harder problem.

Naver runs a comparison-shopping service and at the same time runs an open market called Smart Store.

That is, one company was simultaneously:

the platform providing the search and comparison service

and also

the business owning the sellers competing inside that platform.

That is what it was.

The commission judged that Naver had adjusted its comparison-shopping algorithm several times so that its own Smart Store products were displayed above rival open-market products, and imposed a corrective order and a penalty surcharge of about 26.6 billion won.

The Seoul High Court also upheld the commission's decision.

But here, the 2025 Supreme Court ruling must be read alongside it.

10. The Supreme Court reversed and remanded the Naver case

On October 16, 2025, in Case No. 2023Du32709, the Supreme Court reversed the Seoul High Court's judgment and remanded the case.

The Supreme Court's core point was that the mere existence of self-preferencing by a platform does not by itself constitute abuse of market dominance or an unfair practice, and that the commission must fully prove the legal requirements, including the anticompetitive effect and the intent and purpose behind it.

So the accurate statement at present is:

“Naver's algorithm manipulation was finally confirmed by the Supreme Court.”

is not it.

The accurate statement is:

“The Fair Trade Commission found and sanctioned Naver's algorithmic self-preferencing, and the Seoul High Court agreed, but the Supreme Court held that the commission's legal reasoning and proof were insufficient and reversed and remanded the lower court's judgment.”

That is the correct form.

This difference matters a great deal.

But at the same time, the Supreme Court did not say **“there was no such fact as Naver changing its algorithm.”**

The Supreme Court's own judgment deals with a case in which Naver modified and changed its search algorithm so that its own Smart Store products were displayed above rival open-market products. The issue was the requirements and proof needed for that to count as unlawful self-preferencing under competition law.

In other words, rather than eliminating the question of Naver's power, this case left an even more important question.

«When a platform is both referee and player, how far counts as normal algorithm optimization, and from where does it become anticompetitive self-preferencing?»

In the AI era, this problem grows far larger.

11. Problems arose for consumers, too

In 2025, the Fair Trade Commission also issued a corrective order regarding advertising for Naver Plus Membership.

The commission judged that in advertising membership reward benefits, Naver had failed to adequately disclose the actual restrictions, producing advertising likely to lead consumers to perceive the benefits as larger than they really were. A corrective order was issued not to run similar misleading advertising in the future.

This one case does not define Naver's overall corporate ethics.

But seen together with the earlier cases, a pattern emerges.

Toward workers, organizational power;

toward businesses, platform power;

toward competitors, informational and exposure power;

toward consumers, information-asymmetry power.

Naver holds different forms of power in different relationships.

12. So the phrase “anti-human-rights” can be moved from emotion to structure

“Naver is an anti-human-rights company.”

In a legal sense, this sentence is far too broad.

But rewrite it as follows, and it becomes a powerful proposition.

«Naver is a company that, in the course of exercising the organizational, market, and informational power held by a giant platform, has repeatedly experienced real problems and sanctions concerning labor rights, organizational culture, competitive order, and consumer information asymmetry.»

This is not mere emotion.

The labor problems were confirmed by a government investigation: the Ministry of Employment and Labor's special inspection.

The exclusive handling of real-estate information was confirmed by a Fair Trade Commission sanction.

The shopping-algorithm problem went through a judicial process: commission sanction → Seoul High Court affirmation → Supreme Court reversal and remand.

The membership-advertising problem likewise led to a commission corrective order.

And inside the organization, in 2025, 98.82% of the union voted against the return of an executive.

It is hard to explain all of this as “the deviation of one or two people.”

13. But what matters more is Naver's product experience

Here, let us leave law and regulation and return to the user's experience.

The reason Naver was strong in the past was that it had an enormous number of functions.

Search,

news,

blogs,

cafés,

shopping,

maps,

reservations,

Pay,

content,

advertising,

community.

Naver bundled the many functions of the internet into one place.

In the early days of the internet, this was an enormous advantage.

But once the internet matures, this structure can work in reverse.

Having many functions is not necessarily good UX.

If what the user wants is a single thing,

then having countless services linked on one screen is not convenience but cognitive cost.

And AI solves this problem in exactly the opposite way.

Users can ask a question without reading search results one by one.

They can explain context to the AI.

They can ask follow-up questions.

They can have multiple sources synthesized.

That is:

the Naver-style interface

«keyword → list of results → click → compare → judge»

gives way to

the AI-style interface

«problem → conversation → synthesis → judgment»

That is the migration.

This is not a simple UI change.

It is a change of cognitive interface.

14. From Kahneman's perspective, this is a migration of habit

People do not objectively evaluate a platform every time.

They repeatedly use the path that gives them the result they want with the least effort.

At first:

they use Naver and also try AI.

Next:

they use AI first and Naver only when they need it.

Finally:

they don't think of Naver at all.

Here, the most dangerous thing for a platform is not cancellation.

It is being forgotten.

Users leaving because they dislike Naver matters, but what is far more frightening is users no longer thinking of Naver as an option in the first place.

At that moment, the network effect begins to run in reverse.

15. From Hermann Simon's perspective, the problem is pricing power

The economic value of a search engine does not lie in the number of searches itself.

It lies in how many of the user's decision-making moments it can capture and convert into economic value.

Search:

→ advertising

→ click

→ transaction

→ purchase.

The stronger this chain, the higher the price it can charge advertisers.

But if users begin their important decisions in AI, the story changes.

When AI synthesizes information and recommends a product or service,

the user no longer necessarily has to pass through Naver's search ads.

Then the economic value of the search traffic Naver holds changes.

Traffic remaining and pricing power remaining are two different things.

The real question of the AI era is:

«“How many times do people visit Naver?”»

is not it, but rather

«“Of the important economic decisions being made, how much value is Naver capturing?”»

That is the question.

16. From Herb Cohen's perspective, the problem is Power

For Cohen, power is not simply a title or money.

What matters is how many options the other side has.

This is also why Naver was strong.

Sellers had to meet customers on Naver.

Advertisers had to secure exposure on Naver.

Content producers had to get search traffic from Naver.

Consumers looked for information on Naver.

As this repeated, dependence → bargaining power was created.

But what happens when users and businesses gain many substitutes?

Google.

YouTube.

Instagram.

ChatGPT.

Gemini.

Specialized services.

Direct search.

All kinds of commerce platforms.

The easier multi-homing becomes, the more Naver's bargaining power falls.

So the real crisis for a platform is not the moment its users vanish all at once.

It is the moment the thought “it doesn't have to be Naver” spreads across every economic actor.

17. From Don Norman's perspective, an even more fundamental change is happening

The Naver era was the era of the portal interface.

Everything was gathered in one place, and users had to find and enter each function.

But the AI era is different.

Instead of hunting through services, the user states a goal.

Search.

Compare.

Summarize.

Analyze.

Plan.

Execute.

All of these steps connect within a single conversational interface.

From Norman's perspective, this is not an increase in functions but a change in the mode of behavior itself.

And if this change continues, Naver's problem is not “how much can we improve search quality?”

It becomes whether the search box, as an interface, can remain the most important middle layer between humans and the internet.

18. So Naver's real competitor is not Google alone

This is the most important conclusion.

Thinking only about Naver losing search share to Google is seeing the problem far too small.

It is true that in 2026 the Google app's monthly active users in Korea surpassed Naver's, but that alone cannot be taken to mean a reversal of the entire search market.

The bigger change is this.

Portal

↓

Search

↓

AI

↓

Agent

That is the migration.

Google can be Naver's competitor and at the same time become the platform for AI.

ChatGPT and Gemini can be competitors to search engines and at the same time become interfaces that bypass search itself.

And the barriers to Google Maps entering Korea have weakened, too.

In February 2026, the Korean government conditionally approved the export of high-precision map data that it had restricted for about 20 years. Google must comply with several conditions, including keeping security facilities undisclosed and processing on domestic servers. This decision opened an institutional path for Google Maps to develop into a far more complete service in Korea than before.

In other words, maps — one of Naver's last strong points of contact with daily life — also enter into competition over the long term.

19. Then what should “Naver is going under” actually mean?

Predicting that Naver will necessarily go bankrupt or be delisted by 2030 is a stretch on current evidence.

As of 2026, Naver is still a large company with strong cash-generating power, and its revenue is growing. In the second quarter of 2026, revenue rose 16.2%.

Therefore

“Naver is about to collapse.”

is not analysis.

However,

“Naver may lose its status as the default interface of the Korean internet.”

is a thoroughly realistic analysis.

And these two events are completely different.

When Internet Explorer disappeared, Microsoft did not disappear.

When BlackBerry collapsed, the company did not vanish immediately either.

Historically, the death of a platform often comes before the death of the company.

Stepping down from being the social default.

That is the historical death of a platform.

20. Naver's real danger is not “user exodus” but “the disappearance of necessity”

Naver's strongest line of defense was habit.

“Internet = Naver.”

The next line of defense was the ecosystem.

“Search on Naver, shop on Naver, maps on Naver, content on Naver.”

But AI and specialized platforms are beginning to disassemble these functions one by one.

Information: AI.

Original web sources: Google.

Video: YouTube.

Shopping: specialized commerce.

Information on people and companies: specialized intelligence services.

Maps: competition between Google Maps and Naver Map.

Payments: separate financial services.

Content: each on its own platform.

Then the core value of the portal is reversed.

From “everything is in one place”

to

“use the best tool for each purpose.”

That is the shift.

At that point, the complexity Naver holds can become a cost rather than an asset.

21. Naver's paradox

Naver is actively adopting AI right now.

In fact, the company is integrating AI into core businesses such as advertising and commerce in 2026, and it explains that AI integration contributed to second-quarter revenue growth. At the same time, partly because of investment in AI infrastructure, operating profit fell 0.2% year on year.

This is not a story of Naver foolishly sitting on its hands.

Quite the opposite.

Naver knows the problem and is responding.

But there is no guarantee that this solves the structural problem.

Because Naver adopting AI

and users choosing Naver as their default AI interface

are two different problems.

A platform's most important asset is not the technology itself.

It is the user's habit.

22. In the end, Naver's history is also a history of power

Naver made the Korean internet enormously convenient.

There is no need to deny that fact either.

But as convenience grew, so did dependence.

As dependence grew, so did the platform's power.

And as power grows, responsibility must grow with it.

Toward workers, there is a responsibility to protect.

Toward businesses, a responsibility to apply fair rules.

Toward competitors, a responsibility not to unfairly block market access.

Toward consumers, a responsibility to provide accurate information.

And a company that influences people's environment of choice through algorithms must bear transparency and accountability commensurate with that algorithmic power.

The 2021 special labor inspection showed that Naver had exposed serious problems on the internal side of this responsibility.

The real-estate case and the shopping-algorithm case showed how large the platform's power in the external market was.

The 2025 controversy over Choi In-hyuk's return and the union's 98.82% opposition were a strong signal that this past had not yet been fully resolved inside the organization.

23. So the problem is not Naver alone

Nor does the problem automatically get solved if Naver disappears.

If Google seizes everything, another power arises.

If an AI company becomes the gateway to all information, another power arises.

The problem of platforms is not the nationality of a particular company.

It is who monopolizes the gateway to information, transactions, and judgment.

So the future we want

is not Naver → Google,

yet another replacement of one monopoly by another.

The real future is:

Naver / Google / AI / specialized platforms / direct transactions

competing with one another,

users moving freely,

sources of information transparent,

businesses not captive to a single platform,

workers not silenced by organizational power,

and a structure in which, when an algorithm changes the human environment of choice, it bears responsibility commensurate with that influence.

24. This is the real end of the Naver era

Naver does not need to shut down.

It does not need to be delisted.

Its headquarters does not need to disappear.

The real change is far simpler.

One day, when Koreans want to know something, they no longer automatically think of Naver.

When someone researches a company, they ask an AI instead of Naver.

When comparing something, they look at an AI's analysis instead of Naver's search results.

When checking an original source, they open Google.

When watching video, they open YouTube.

When researching people and companies, they use specialized intelligence services.

When finding their way, they choose the best map.

And nobody thinks of any of this as unusual behavior.

That is when Naver's historical standing ends.

It is not the company that dies. It is the era.

25. Final judgment

As of September 2026, there is no basis for assessing Naver as a company on the verge of bankruptcy.

On the contrary, financially it is strong.

But there is considerable basis for criticizing Naver's legitimacy and structural position.

On the labor side, the 2021 government special inspection confirmed workplace harassment, problems in the reporting and protection system, and violations of labor law.

On the market side, Naver was sanctioned by the Fair Trade Commission for the exclusive handling of real-estate information, and the shopping-algorithm self-preferencing case went through lengthy proceedings before the commission and the courts. However, since the Supreme Court reversed and remanded the shopping case in 2025, it should not be stated that Naver's illegality has been finally confirmed.

On the consumer side, too, the commission issued a corrective order in 2025 regarding Naver Plus Membership advertising.

On the side of organizational trust, 98.82% of participants in Naver's union vote opposed the 2025 return of former COO Choi In-hyuk.

And on the side of user behavior, in 2026 the Google app's monthly active users in Korea surpassed Naver's for the first time, and with the conditional easing in 2026 of restrictions on Google Maps' access to Korean map data, an environment has formed in which Naver's existing status as a platform for daily life is exposed to competition over the long term.

So the most accurate conclusion is this.

«Naver's problem is not simply that its search quality is low.

For a long time Naver was the giant gateway connecting Koreans' information, transactions, and choices, and in that process real problems concerning labor rights, market order, and information asymmetry arose again and again. At the same time, with the arrival of AI and specialized platforms, the very necessity of that giant gateway is weakening.

Therefore Naver's greatest threat is not Google. It is the arrival of a world in which people can make better decisions, use better services, and get better information without Naver.»

From Kahneman's perspective, the path of human judgment migrates.

From Simon's perspective, the economic value that can be captured from that judgment migrates.

From Cohen's perspective, dependence on the platform weakens and bargaining power migrates.

From Norman's perspective, the search-box-centered interface itself is replaced by conversational, agent-style interfaces.

And what decides Naver's historical fate, in the end, is the intersection of these four.

When the people leave, the money follows, the dependence vanishes, and the interface changes, the era of a platform ends.

It is not a question of whether Naver can survive.

It is a question of whether Naver is still necessary.

And as of 2026,
for the first time in Korea, the answer to that question
has begun to change to “not necessarily.”