South Korean internet company Naver has launched HyperCLOVA X, a generative AI platform that pairs a conversational chatbot with a planned AI-powered search service. The rollout puts Naver’s own large language model at the center of its effort to expand beyond its existing search offering.
A chatbot launches as search plans take shape
Naver Cloud said CLOVA X began beta service in English and Korean on August 24. The chatbot is designed to offer a ChatGPT-like experience, while Cue is intended to bring generative AI into search.
Cue was expected to become available for beta testing in September, with plans for it to be incorporated into Naver’s current search engine in November. That sequence gives the company a path from a standalone conversational service toward AI features within a product people already use for search.
The two offerings have different roles in Naver’s rollout. CLOVA X is the conversational service, while Cue is the search-focused product. Both sit within the HyperCLOVA X effort, which Naver presents as a platform for creators and enterprise customers as well as a foundation for its own services.
Building on a Korean-language model
HyperCLOVA X is an upgraded version of HyperCLOVA, Naver’s earlier Korean-language large language model launched in 2021. Naver says HyperCLOVA has more than 204 billion parameters, but the company did not disclose how many parameters HyperCLOVA X was trained on.
The distinction matters when describing the launch: the parameter figure belongs to HyperCLOVA, not a disclosed specification for the new model. Naver also said users must sign up for Naver accounts to use HyperCLOVA X, so access is tied to the company’s account system.
Naver’s leadership has emphasized the company’s AI workforce and technical ambitions. CEO SooYeon Choi said in a letter to shareholders published earlier this week that Naver has more than 500 AI experts and is one of five companies worldwide to have developed its own large language model with more than 100 billion parameters.
More than text is on the roadmap
Naver Cloud is also developing multimodal language models for CLOVA X. At the company’s conference on Thursday, Nako Sung, head of Technology and Hyperscale AI at Naver Cloud, said those models would be capable of analyzing and generating images, videos and audio as well as text.
Sung said Naver aims to release the multimodal model in the near future. The announcement describes a development goal, rather than a feature already available in the beta service. For now, the launch centers on the chatbot and the upcoming search test, while work continues on broader kinds of input and output.
The company is also investing in infrastructure for its generative AI services. Choi said Naver plans to open its second data center, GAK Sejong, in South Korea in November. The center is expected to have 600,000 servers. Naver has also been working with Samsung to develop AI chips for hyperscale AI since last December.
Competing through regional focus
Naver’s announcement comes as technology companies compete to build and extend generative AI products. The source describes activity from OpenAI, Google and Baidu, while Meta and Amazon were reportedly preparing their own AI-powered chatbots. That competitive backdrop makes Naver’s combination of a homegrown model, chatbot and search integration part of a wider push across the industry.
Naver says it is looking toward regions competitors have not yet covered. It has discussed tailored AI applications for countries with political sensitivities in the Middle East and for non-English-speaking countries such as Spain and Mexico, alongside South Korea, Japan and Southeast Asia.
The strategy connects the technical rollout to potential markets: language and local context are central to the company’s stated plans. HyperCLOVA X begins with English and Korean beta access, while Naver describes a broader regional ambition. Whether that direction distinguishes its services will depend on how the chatbot, search integration and future multimodal capabilities develop.