[Edaily Kim Hyun-ah Reporter] While network traffic is skyrocketing, telecom companies’ profitability has been slow to keep pace. In the 5G era, data usage has surged due to the proliferation of OTT services, and the three major domestic telecom companies have invested approximately 50 trillion won in network infrastructure, including 5G, to date. However, despite these massive investments, profits have not increased significantly.

According to the telecommunications industry, revenue for the three major domestic telecom companies has increased approximately fourfold, from 13 trillion won in 2001 to 60 trillion won in 2025. In contrast, operating profit during the same period rose only from 3.3 trillion won to the 4 trillion won range. While revenue has grown significantly, profits have remained stuck in a range between 2 trillion and 4 trillion won.

The problem is that another massive surge in traffic is expected in the AI era.

Unlike existing OTT services, generative AI involves users and AI exchanging data through repeated question-and-answer interactions. As AI agents become more widespread, there is also the possibility that they will continuously communicate with various services on behalf of users. If this expands to physical AI—such as autonomous driving and robotics—demands on network quality, including not only speed but also stability and ultra-low latency, will increase.

If telecommunications carriers bear the costs of building and maintaining networks but the increase in traffic does not translate into revenue and profits, they risk remaining mere “dumb pipes” that simply transmit data.

Lee Jong-kwan, Senior Research Fellow at Sejong Law Firm, pointed out the structural problems facing the telecommunications industry during a seminar held on the 17th at Korea University’s Center for Technology Law and Policy, citing the European Union’s (EU) “Digital Network Act (DNA)” as an example of a new network policy.

50 Trillion Poured In… But Profits Stagnant

The problem facing the telecommunications industry is not that traffic isn’t growing. The issue lies in the fact that while traffic has continued to increase, it has been difficult to convert this into revenue.

“The costs of deploying networks have become decoupled from revenue,” said Lee. “With expected future returns at a significantly low level, companies are inevitably left wondering whether they should invest at all.”

The three major domestic mobile carriers have poured massive funds into building communication networks, including 5G. Cumulative capital expenditures have reached approximately 50 trillion won.

However, profits have not increased in proportion to the investment. Revenue for the three telecom companies, which stood at 13 trillion won in 2001, increased approximately fourfold to 60 trillion won in 2025. In contrast, operating profit rose only from 3.3 trillion won to around 4 trillion won.

Critics point out that their profitability remains low even when compared to global telecommunications companies. While Verizon in the U.S. has a five-year average operating profit margin of about 22% and Japan’s KDDI stands at around 18%, the operating profit margin of South Korean telecommunications companies is only about 7.1%.

From the telecom companies’ perspective, this means they remain stuck in a structure where, even as traffic increases and network investments grow, it is difficult to recoup these costs through sufficient profits.

Under these circumstances, if AI generates a new surge in traffic, the investment burden on telecom companies could increase once again.

In the AI Era: More Traffic, for Longer Periods

OTT services were the primary drivers of network traffic growth during the 5G era.

As users streamed videos, massive amounts of data were delivered to them via telecommunications networks. The direction of traffic growth was also relatively clear: it was a “heavy download” model.

In the AI era, the pattern may change.

With generative AI, users enter a question, the AI provides an answer, and the process repeats as the user asks another question. Unlike a search, where the process ends after viewing a single result, interaction with the network continues—such as uploading data and refining results.

Commissioner Lee Jong-kwan stated, “Following the one-way download load caused by OTT, the AI era will mark ‘Traffic Surge Season 2,’ where the volume of uploaded data itself increases.”

There is also a possibility that AI services will occupy network bandwidth for longer periods. Commissioner Lee pointed out that AI services “occupy the network for an extremely long time.”

These changes could become even more significant as AI agents become more widespread. This is because AI can search on behalf of users, exchange information with other AI systems, and connect various services such as shopping, booking, and business systems. This involves AI exchanging data with multiple services in the background without the user having to click anything directly.

From the perspective of telecom carriers, as traffic increases, investment is needed to enhance network capacity and quality. However, if the existing flat-rate pricing structure remains in place, it is difficult to view the increase in traffic as directly translating into additional revenue.

This suggests that the problem of “traffic increasing but revenue not keeping pace,” which characterized the 5G era, could repeat itself in the AI era.

Europe Boosts Telecom Operators’ Bargaining Power with ‘DNA’

The EU is promoting the “Digital Networks Act (DNA)” to address structural issues in the telecommunications industry.

The main objective is to consolidate the telecommunications markets, which are currently fragmented across individual countries, into a single market.

Commissioner Lee explained the essence of the DNA as “moving from the autonomy of individual nations to a supranational single market.”

The vision is to increase the competitiveness of European telecom companies by expanding the market size and to strengthen their bargaining power against global Big Tech companies.

The DNA includes a “Single Passport” system that allows companies to operate throughout the entire EU with a single license obtained in one country.

It also includes measures to increase the stability of spectrum licenses, thereby encouraging long-term investment by telecommunications companies.

Mechanisms will also be established to allow regulatory agencies to intervene in the event of disputes between telecom operators and big tech companies regarding network access.

Commissioner Lee explained this as “a mechanism to balance bargaining power so that unilateral, unfair situations do not arise.”

While maintaining the principle of net neutrality—which prohibits traffic discrimination—the plan will also institutionalize “special services” that allow for separate traffic management for innovative services.

The AI Era Calls for Customized Networks and Billing

In

the

AI

era,

the role of telecommunications networks may also change.

Moving beyond the practice of transmitting all data in the same way, there is a growing need to differentiate networks based on the quality and speed required for specific services.

In particular, for physical AI applications such as autonomous driving and robotics, stable communication and ultra-low-latency networks are critical.

Commissioner Lee stated, “In the era of 6G and AI, a B2B market far more expansive than the existing 5G market will open up,” adding, “In the case of physical AI, which requires high precision or ultra-low latency, it must effectively be classified as a special service.”

For example, services where communication latency is critical—such as robots in factories or autonomous driving systems—require a level of network quality distinct from that of general consumer internet.

For telecom carriers, such services have the potential to become new revenue streams.

This marks a shift away from a pricing model based solely on “how much data was used” toward one based on network quality—such as speed, stability, and latency.

This means that while AI places new demands on communication networks, it can also open up a new B2B market for telecom carriers.

“Beyond Infrastructure Development: Toward ‘Competition Policy’”

Some argue that the direction of telecommunications policy must also change.

Until now, government policy has focused on expanding telecommunications infrastructure and increasing investment in facilities. However, when network investment does not translate into actual profits, simply expanding investment has its limits.

Commissioner Lee stated, “While the past approach centered on ‘industrial policy’—raising the capital-to-equipment ratio through infrastructure investment—we now need to adopt a ‘competition policy’ perspective to level the playing field (with OTT services and others), given that network operators’ influence and competitiveness are waning.”

The challenge for telecommunications companies in the AI era goes beyond simply building more networks. As the industry shifts from OTT to AI, and from AI to agents and physical AI, telecommunications networks are likely to be used more extensively, for longer periods, and with greater sophistication.

The key question is how to convert this growing traffic into revenue. A new challenge for the telecommunications industry is whether telecom companies—which have invested 50 trillion won in network infrastructure—will remain confined to the role of data transmitters while shouldering investment costs in the AI era, or whether they will be able to transform traffic into revenue by leveraging their high-quality networks as a new AX platform business.