South Korea’s Top Tech Players Pledge $390 Million in AI Initiative to Compete with Global Giants
South Korea’s tech sector is making significant strides in the development of large language models tailored to its unique linguistic and cultural landscape. The nation recently announced an ambitious AI initiative, allocating ₩530 billion (approximately $390 million) to five local companies building large-scale foundational models, with the aim to compete on a global level against powerhouses like OpenAI and Google.
This move signifies Seoul’s intention to reduce dependence on foreign AI technologies and strengthen its national security by maintaining tighter control over data in the AI era. The chosen organizations for this undertaking were LG AI Research, SK Telecom, Naver Cloud, NC AI, and start-up Upstage.
Over a six-month review period, the government will assess the progress of the initial cohort, cut underperforming companies, and continue supporting the frontrunners until only two remain to spearhead South Korea’s sovereign AI drive.
Each organization brings distinct advantages to South Korea’s AI race. LG AI Research, a research division of South Korean conglomerate LG Group, offers Exaone 4.0 – a hybrid reasoning AI model that combines broad language processing with advanced reasoning features initially introduced in the company’s earlier Exaone Deep model.
Exaone 4.0 (32B) performs competitively against other models on Artificial Analysis’s Intelligence Index benchmark, with Upstage’s Solar Pro2 also scoring well. To improve and climb the ranks, LG is focusing on refining data before feeding it to the models for training.
Instead of prioritizing sheer scale, LG aims to make the entire process more intelligent, delivering real-world value that surpasses what general-purpose models can offer. This strategy, as co-head Honglak Lee explained to our reporters, is integral to LG’s approach.
LG intends to enhance its models by offering them through APIs and utilizing the data generated by users of those services to train the model for improvement. According to Lee, this partnership results in better services that generate more economic value and richer data.
However, rather than investing in massive GPU clusters, LG AI Research is focusing on efficiency, maximizing output from each chip, and creating industry-specific models. The goal is not to outspend the global giants but to outsmart them with high-performing yet more efficient AI solutions.
South Korea’s telecom giant SK Telecom (SKT) launched its personal AI agent A. service in late 2023, followed by the release of a new large language model, A.X, in July this year. Built on Alibaba Cloud’s Qwen 2.5 open-source model, A.X 4.0 comes in two versions: a 72-billion-parameter model and a lighter 7B version.
SK reports that A.X 4.0 processes Korean inputs 33% more efficiently than GPT-4o, highlighting its local language advantage (OpenAI’s GPT 5.0 comparison data is currently unavailable). SKT also open-sourced its A.X 3.1 models earlier this summer. Meanwhile, the A. service offers features like AI call summaries and auto-generated notes, boasting over 10 million subscribers as of August 2025.
SK’s strength lies in versatility due to its access to information from its telecom network, ranging from navigation to taxi-hailing services. Taeyoon Kim, head of the foundation model office at SK Telecom, explained that the company aims to bridge the gap between cutting-edge model research and real-world impact by leveraging its telecom infrastructure, extensive user base, and proven service like A.
SK Telecom is also investing in AI infrastructure, using GPUaaS (South Korea’s largest GPU-based service) and building a new hyperscale AI data center with AWS. To address any gaps, the company partners with Korean AI chipmaker Rebellions, secures trusted data partnerships through government and university collaborations, and fosters a global research network.
Naver Cloud, the cloud services arm of South Korea’s leading internet company, introduced its large language model, HyperClova, in 2021. Two years later, it released an upgraded version, HyperCLOVA X, as well as 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.
Naver Cloud views large language models (LLMs) as connectors linking legacy systems and siloed services to improve usefulness. Naver stands out as Korea’s only company, and one of the few globally, that can genuinely claim an “AI full stack.” It built its HyperCLOVA X model from scratch and manages the massive data centers, cloud services, AI platforms, applications, and consumer services that bring the technology to life.
Similar to Google but tailored for South Korea, Naver is embedding its AI into core services like search, shopping, maps, and finance. Its advantage lies in real-world data, with AI Shopping Guide offering recommendations based on actual purchasing preferences, among other services.
To compete with global AI giants like OpenAI and Google, Naver prioritizes perfecting its “recipe” for models and securing the capital to scale them. However, instead of focusing solely on size, the company emphasizes sophistication, arguing that its AI is already globally competitive at comparable scales.
Upstage is the only start-up participating in the project. Its Solar Pro 2 model, launched last July, was recognized as a frontier model by Artificial Analysis, placing it among OpenAI, Google, Meta, and Anthropic (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 exceptionally well for South Koreans and is more cost-effective, according to Kwon.
Upstage aims to distinguish itself by focusing on real business impact rather than just benchmarks. To this end, the company is developing specialized models for industries like finance, law, and medicine while striving to build a Korean AI ecosystem led by “AI-native” startups.
