48 Billion Hours Spent on Generative AI Last Year
Changing User Engagement Signals a New Phase in AI Competition
Longer User Retention and More Diversified Subscription Plans

Last year, global users spent a total of 4.8 billion hours using generative artificial intelligence (AI) applications, such as OpenAI's ChatGPT and Google's Gemini. As usage time increased by 3.6 times compared to the previous year, global big tech companies that had previously focused on user acquisition are now shifting their focus to maximizing profits, intensifying competition in the process.

AI applications such as DeepSeek, ChatGPT, and Gemini displayed on a mobile phone screen. Photo by Yonhap News

AI applications such as DeepSeek, ChatGPT, and Gemini displayed on a mobile phone screen. Photo by Yonhap News

View original image


According to global app market analytics firm Sensor Tower on February 3, the total global consumption time for generative AI applications reached 4.8 billion hours last year, a 3.6-fold increase from 1.34 billion hours in 2024. This figure is about ten times higher than in 2023.


The number of generative AI app downloads reached 3.8 billion, double the previous year, while in-app purchase revenue surged threefold, surpassing 5 billion dollars (approximately 7 trillion won). The number of sessions (one-time app visits) exceeded 1 trillion. Notably, the year-on-year session growth rate (256.4%) outpaced the download growth rate (117.7%), indicating a shift in generative AI app trends from acquiring new users to strengthening engagement among existing users.


As user engagement patterns change, AI competition has entered a new phase. Big tech companies, which have made massive investments in AI technology and expanded its use cases to compete with ChatGPT, have now entered a phase of accelerating profits and are engaging in price competition. The competition is intensifying through the launch of subscription models, diversification of pricing plans, and efforts to increase user retention time.


Google and OpenAI are targeting users by segmenting their pricing plans. Recently, Google launched the 'AI Plus' plan, priced at 11,000 won per month, in 35 countries and regions including Korea and the United States. This is a budget-friendly plan costing less than half of the existing 'AI Pro' plan (29,000 won per month). It targets customers who want better performance than the free plan but find the existing plan burdensome. In practice, users can access higher-tier AI models such as 'Gemini 3 Pro' and 'Nanobanana Pro' more frequently than free users. However, credits (usage limits) and cloud storage capacity are provided in smaller amounts compared to the existing plan.

Generative AI Soaks Up 4.8 Billion Hours in a Year... Profit Race Heats Up as It Becomes a "Viable Business" View original image

OpenAI took the lead over Google by introducing the low-cost 'ChatGPT Go' plan, expanding it from select countries to a global scale. At 8 dollars per month (15,000 won in Korea), users can utilize AI more extensively than with the free version and access core features such as image generation and file uploads. Advertisements are displayed based on the information users seek.


Meta expects this year to be when AI directly drives product and business performance. For example, Meta is developing a system where AI manages the entire advertising process from creation to optimization, requiring advertisers to simply input product links. Like Google and OpenAI, Meta is enhancing AI features across all its services, including Facebook, Instagram, and WhatsApp, and is preparing to introduce paid subscription services. AI-based content recommendation and advertising systems are expected to increase user retention time and boost revenue.



An industry insider commented, "Although there are claims of an AI bubble, performance is gradually improving in line with investments, and big tech companies are focusing on strategies to build sustainable business models. As the paradigm shift toward AI agents has already begun, the industry will increasingly be reorganized around AI cost and operational efficiency, infrastructure, and execution capabilities."


This content was produced with the assistance of AI translation services.

© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.

Today’s Briefing