From technology titans to emerging companies, South Korean participants are creating extensive language frameworks customized to their native language and culture, prepared to challenge international heavyweights like OpenAI and Google.
Last month, the nation initiated its most daring sovereign AI campaign to date, committing ₩530 billion, about $390 million, to five domestic firms constructing large-scale foundational frameworks.
The initiative highlights Seoul’s intention to reduce dependence on foreign AI solutions, aiming to enhance national protection and maintain stricter control over information in the AI era. The institutions selected by the Ministry of Science and ICT to compete included LG AI Research, SK Telecom, Naver Cloud, NC AI, and the startup Upstage.
Every half-year, the government will assess the first cohort’s development, remove underachievers, and continue financing the leaders until only two remain to steer the country’s sovereign AI mission. Each participant brings a unique benefit to South Korea’s AI competition. TechCrunch interviewed several of the chosen companies regarding how they plan to confront OpenAI, Google, Anthropic, and others on their home ground. NC AI declined to provide comments.
LG AI Research And Exaone 4.0
LG AI Research, the R&D branch of South Korean conglomerate LG Group, provides Exaone 4.0, a hybrid reasoning AI framework. The current version integrates broad language interpretation with the sophisticated reasoning capabilities first introduced in the firm’s previous Exaone Deep model.
Exaone 4.0 (32B) already performs reasonably well against rivals on Artificial Analysis’s Intelligence Index benchmark, similar to Upstage’s Solar Pro2. But it intends to advance and climb the ranks through its deep access to real-world industry information ranging from biotechnology to advanced materials and manufacturing.
Stay up to date with the latest technology in TheTechCrunch.info, which covers artificial intelligence, mobile and web apps, modern things, cyber security, and general technical news. From AI’s successes to chat and generative tools, such as smartphones, laptops, and wearables’ special reviews, TheTechCrunch gives an insight into this case.
It combines that information with a focus on refining the data before supplying it to the models for training. Rather than pursuing sheer magnitude, LG aims to make the entire process more intelligent, so its OpenAI can provide genuine, practical value beyond what general-purpose models can deliver. This is our fundamental approach, co-head Honglak Lee told TechCrunch.
LG is enhancing its frameworks via proven methods: distributing them through APIs, then leveraging the real-world data generated by users of those services to train the model for improvement. As LG’s models advance, our partners can offer superior services, which subsequently produce greater economic value and even richer information, he added.
Instead of targeting massive GPU clusters, LG AI Research emphasizes efficiency, optimizing every chip, and creating industry-specific frameworks, he mentioned. The objective isn’t to outspend global giants but to outthink them with high-performing, yet more efficient, AI.
SK Telecom’s Personal AI And A.X Model
South Korea’s telecom leader SK Telecom launched its personal AI agent A., pronounced A-dot service, in late 2023 and recently introduced its new large language framework, A.X, this July. Built on the Chinese open-source model from Alibaba Cloud, Qwen 2.5, A.X 4.0 comes in two variants, a hefty 72-billion-parameter version and a lighter 7B version.
SK claims that A.X 4.0 processes Korean inputs around 33% more efficiently than GPT-4o, highlighting its local language advantage. SKT also open-sourced its A.X. 3.1 models earlier this summer. Meanwhile, the A. service offers functionalities like OpenAI call summaries and automatically generated notes. As of August 2025, it has attracted approximately 10 million subscribers.
SK’s strength is its adaptability, because it has access to data from its telecom network, ranging from navigation to taxi-hailing. SK Telecom’s role is to act as a bridge between advanced model research and real-world impact.
With our telecom infrastructure, extensive user base, and proven service like OpenAI, we bring AI directly into daily life, whether in customer service, mobility, or manufacturing, Taeyoon Kim, head of the foundation model office at SK Telecom, told TechCrunch.
SK Telecom is also investing in AI infrastructure, using GPUaaS, South Korea’s largest GPU-based service, and constructing a new hyperscale AI information center with AWS. Whatever it lacks, it is collaborating to obtain.
We’re building a full-stack ecosystem with Korean OpenAI chipmaker Rebellions, securing trusted data partnerships through government and university collaborations, and fostering a global research network, Kim added. That includes projects like our collaboration with MIT, which applies foundational frameworks to advanced manufacturing, battery, and semiconductor innovation.
Naver Cloud And HyperCLOVA X
Naver Cloud, the cloud services division of South Korea’s leading internet company, unveiled its large language framework, HyperClova, in 2021. Two years later, it introduced an upgraded version, HyperCLOVA X, along with new products powered by the technology.
CLOVA X, an AI chatbot, and Cue, a generative AI-driven search engine positioned as a rival to Microsoft’s CoPilot-enhanced Bing and Google’s AI Overview. It also released this year its multimodal reasoning AI framework, HyperCLOVE X Think.
Naver Cloud believes the real strength of LLMs is to serve as connectors linking legacy systems and siloed services to improve efficiency, according to a Naver spokesperson. Naver stands out as Korea’s only firm and one of the few globally that can truly claim to have an OpenAI full stack.
It developed HyperCLOVA X from scratch and operates the massive data centers, cloud services, AI platforms, applications, and consumer services that bring the technology to life, the spokesperson explained.
Explore a complete hub for the latest apps, smart things, and security updates online, ranging from AI-operated solutions and automation tools. TheTechCrunch.info offers in-depth articles, comparisons, and specialist analysis designed to understand the rapidly changing technology. Whether you are keen on robotics, data protection, or the latest digital trends.
Similar to Google but adapted for South Korea, Naver embeds its AI into core services like search, shopping, maps, and finance. Its advantage is practical data. Its AI Shopping Guide offers suggestions based on actual consumer behavior. Other services include CLOVA Studio, which enables businesses to build custom generative AI, and CLOVA Carecall, an AI-powered check-in service for seniors living independently.
The Naver spokesperson emphasized that outperforming global AI giants like OpenAI and Google depends on two factors: refining its model “recipe” and securing resources to scale them. Even so, rather than focusing on size, the company emphasizes sophistication, asserting its AI is already competitive globally at similar scales.
Upstage And Solar Pro 2
Upstage is the only startup participating in the project. Its Solar Pro 2 framework, launched last July, was the first Korean model recognized as a frontier model by Artificial Analysis, placing it alongside OpenAI, Google, Meta, and Anthropic, according to Soon-il Kwon, executive vice president at Upstage.
While most frontier models have 100 billion to 200 billion parameters, Solar Pro 2, with just 31 billion, performs better for South Koreans and is more cost-efficient, Kwon told TechCrunch. Solar Pro 2 has outperformed international models on major Korean benchmarks. With this project, Upstage aims to reach a Korean language performance of 105% of the global standard, Kwon said.
Upstage seeks to distinguish itself by concentrating on real business outcomes, not just benchmarks, he explained. It is developing specialized models for industries like finance, law, and medicine, while pushing to create a Korean OpenAI ecosystem led by AI-native startups.