ChatGPT's iron grip in Korea loosens as Gemini quickly gains ground

Home > Business > Tech

ChatGPT's iron grip in Korea loosens as Gemini quickly gains ground

Published: 28 Feb. 2026, 16:09 Updated: 02 Mar. 2026, 12:58

Audio report: written by reporters, read by AI

The Gemini app icon on a smartphone in this illustration taken October 27. [REUTERS]

A worker, identified only by the surname Ma, switches back and forth between Google’s generative AI Gemini and OpenAI’s ChatGPT. But she found herself spending more time on Gemini at work, as its services, from generating images with Nano Banana to organizing her notes with NotebookLM, can be linked and applied directly to her tasks.

“I ask ChatGPT any light questions that I have, but for work-related tasks that require more deliberation, I usually turn to Gemini. It’s more convenient because various Google AI services are interconnected.”

Ma’s case reflects the increasing adoption of Gemini in Korea, with ChatGPT’s dominance in the generative AI market showing signs of strain.

While the number of users turning to ChatGPT has not been rising as quickly as it used to be, Gemini’s monthly active users in Korea reached 123,647 last month, surpassing 100,000 for the first time since its launch, according to the data tech firm IGAWorks. During the same period, ChatGPT’s monthly active users stood at approximately 15.29 million — an overwhelmingly higher amount of total users, but a flatter growth rate, with total monthly active users only increasing 3.3 percent from 13.84 million compared to the previous month.

The ChatGPT logo is seen in this illustration taken on Jan. 22, 2025. [REUTERS/YONHAP]

While ChatGPT still has an overwhelmingly higher number of total users, its monthly active users only increased 3.3 percent from 13.84 million compared to the previous month. Gemini’s monthly active user base, on the other hand, grew to 94,760, marking a 30.5 percent increase.

According to a survey released by the research firm Consumer Insights on Wednesday, 75 percent of 1,600 adults aged 18 to 65 in Korea said they had experience using AI services.

By model, 60 percent of respondents primarily used ChatGPT, while 25 percent mainly used Gemini. However, Gemini showed strong momentum in terms of app installations. Last month, the Gemini app was downloaded by 458,901 accounts, an increase of about 77,000 compared to the previous month. During the same period, ChatGPT logged 776,297 new downloads, down roughly 120,000 from a month earlier.

Industry insiders see Google’s vast ecosystem as the key driver behind Gemini’s growth. The company has been advancing a multimodal strategy by integrating AI across its search engine’s video, text and image services, as well as other everyday services, from managing schedules through integration with Gmail and Google Calendar to generating automatic subtitles on YouTube.

The company is also broadening its role as a comprehensive AI infrastructure provider by integrating AI into its cloud services, semiconductors and software.

A representative example is Ironwood, its seventh-generation AI chip unveiled last year. Ironwood serves as the core hardware for Google Cloud’s AI Hypercomputer system, which combines AI with computing, networking, storage and software to maximize overall performance and efficiency.

Google has even opened its tensor processing unit (TPU) ecosystem to external partners to strengthen its leadership in the AI ecosystem.

The Information, a U.S. information technology outlet, reported on Thursday that Google had signed a multibillion-dollar contract to provide its TPU AI chips to Meta. In addition to the computing lease agreement, Meta is reportedly discussing a plan to install Google’s TPUs in its data centers.

Cooperation between Google and Apple has also been drawing attention. Last year, Google signed a multiyear agreement to embed Google’s technology into Apple’s Foundation Model, a large language model under development for integration into Apple’s ecosystem, such as with the voice assistant Siri. Though Apple planned to introduce AI features to Siri starting in 2024, it has faced development challenges. Industry observers say that integrating next-generation Siri with Gemini could enhance the voice assistant’s personalized services by understanding on-screen context in real time and conducting more natural conversations.

Google's promotional image of the Gemini 3.1 Pro. [SCREEN CAPTURE]

Another pillar of Google’s Gemini strategy is advancing model performance. On Feb. 19, Google unveiled the Gemini 3.1 Pro Preview version, a successor to Gemini 3.1 released in November last year, with significantly improved reasoning capabilities for solving complex problems. The new model shows the ability to visually explain complex topics and synthesize different data into a unified perspective. Gemini 3.1 Pro Preview scored 77.1 percent on ARC-AGI-2, a benchmark that evaluates AI reasoning capability — more than double the score of its predecessor.

OpenAI, meanwhile, has chosen to accelerate the adoption of enterprise AI agents through partnerships with global consulting firms. The move is interpreted as an effort to upgrade its revenue model by embedding AI into core corporate processes rather than simply selling generative AI models.

This article was originally written in Korean and translated by a bilingual reporter with the help of generative AI tools. It was then edited by a native English-speaking editor. All AI-assisted translations are reviewed and refined by our newsroom.
BY SEO JI-WOO [[email protected]]