[Edaily Reporter Lee So-Hyun ] Kakao(035720) Kakao AI, a new entity to be established through Kakao’s corporate spin-off, has announced its goal of achieving over 6 trillion won in revenue and over 1 trillion won in revenue from its artificial intelligence (AI) business by 2030. The plan is to transfer subsidiary management functions to Kakao X and focus its capabilities on AI, advertising, and commerce centered around KakaoTalk to create new revenue streams.

Jeong Shin-ah, CEO-designate of Kakao AI, stated at the Kakao corporate spin-off press briefing on the 21st, “With this spin-off, Kakao AI has established a clear business structure that allows us to focus solely on our core businesses by separating the significant role of managing subsidiaries.”

Moving forward, Kakao AI will enhance its existing advertising and commerce businesses with a focus on AI, while simultaneously cultivating new revenue streams such as agent-based advertising, agent-based commerce, and subscriptions. The company aims for revenue from its newly launched AI initiatives to account for a double-digit percentage of total revenue by 2028, and to generate over 1 trillion won in revenue from the AI business alone by 2030.

Total revenue is projected to grow at an average annual rate of approximately 20% through 2030, reaching over 6 trillion won. For the same period, the company has set targets of over 2 trillion won in earnings before interest, taxes, depreciation, and amortization (EBITDA), an operating profit margin of at least 30%, and a return on equity (ROE) of at least 20%.

CEO Jeong stated, “Kakao AI aims to achieve revenue of over 6 trillion won and EBITDA of over 2 trillion won by 2030 while maintaining an average annual revenue growth rate of around 20%.” He added, “Along with revenue growth, we will significantly improve profitability and capital efficiency to achieve an operating profit margin of over 30% and a return on equity (ROE) of over 20% by 2030.”

KakaoTalk: A Platform for 50 Million ‘Agent AI’

At the heart of Kakao AI’s growth strategy lies KakaoTalk, which has approximately 50 million users. Kakao plans to evolve KakaoTalk into a platform that combines the user relationships and everyday contextual data accumulated through the app with AI to understand user intent and facilitate the execution of necessary services.

CEO Jeong explained, “By integrating on-device AI technology into KakaoTalk—Kakao’s core asset and the touchpoint closest to the daily lives of our 50 million users—we can now realize the latent value of the vast amount of contextual information from everyday life.” He went on to emphasize, “Another powerful competitive advantage of KakaoTalk is that, once user context is captured within a conversation, various services are connected within the app, allowing the entire process—from intent recognition to actual execution—to be completed as a single, seamless flow.”

Within KakaoTalk, specialized agents in various fields are connected based on a general-purpose Agent AI assistant, tailored to the user’s needs. The long-term vision is to implement a “digital twin” that understands the user, makes decisions on their behalf, and utilizes the necessary agents.

CEO Jeong said, “Ultimately, KakaoTalk will go beyond simply conversing with a single Agent AI; users will be able to directly experience the entire process—from problem-solving to completion—as various Agents are organically connected.”

Through this, the goal is to have more than 20 million users utilize Kakao’s AI daily by 2030. As the use of Agent AI expands, the company forecasts that time spent on KakaoTalk will increase by more than 50% compared to current levels.

CEO Jeong said, “In the future, we expect that everyone in the country will have their own agent on KakaoTalk, and the platform will evolve beyond a messenger connecting people to one another to one where interactions between people and Agent AI—and even among Agent AIs themselves—will become more active,” adding, “Our goal is to have more than 20 million users using AI daily by 2030.”

Expanding into Search Advertising… Aiming for 1 Trillion Won in AI Revenue

The company has also outlined concrete plans to convert AI into actual revenue. Kakao AI has identified agent advertising, agent commerce, and subscriptions as the primary revenue streams for its AI business.

In the advertising sector, the company will introduce new ad products where AI identifies user intent to suggest products and services and facilitate purchases. It also plans to enter the search advertising market with the goal of securing a market share of at least 5% by 2030. For Agent AI, Kakao AI will expand integration with external vertical services and increase purchase conversion rates to cultivate transaction fees as a major revenue source. The company will also expand its B2C AI subscription offerings.

CEO Jeong stated, “We plan to further boost the growth potential of advertising and commerce through integration with AI, and to cultivate AI as a new growth engine with its own revenue model.”

The company also unveiled a cost-efficiency strategy to provide AI to 50 million users. Kakao AI is building AI models optimized for agent services—from on-device to server-based—based on its proprietary multi-model and lightweight technologies.

Approximately half of user requests are processed directly on the device using its own lightweight models, while server AI models and specialized agents are utilized only for complex requests. The “AI Conductor,” currently under development, assesses the complexity and difficulty of requests to select and connect users to the appropriate AI model.

Kakao estimates that by applying on-device-first processing and intelligent model routing, it can reduce server inference costs by up to 90% while maintaining performance. The goal is to build a structure where AI inference costs do not increase at the same rate as the number of users, even as the user base grows.

CEO Jeong stated, “A more efficient AI architecture lays the foundation for more users to use AI more frequently,” and outlined a vision to create a virtuous cycle where expanded usage leads to the participation of various agents and partners, the accumulation of contextual data, and improved AI experiences.