South Korea is preparing a much larger public push into artificial intelligence, with a plan centered on building a frontier AI model developed at home. The proposal would use government-backed equity investments of 4.7 trillion won ($3.49 billion) and place domestic AI capability at the center of the country’s technology strategy.
The money is tied to the proposed 2027 budget, so it is not final yet. Parliament still needs to approve the funding before the program can move from plan to execution.
A bigger bet on domestic AI capability
The core of the proposal is straightforward: South Korea wants a homegrown frontier model. That means the government is not only looking at AI as a service to buy from global providers, but as strategic infrastructure that Korean companies should be able to build and operate.
The planned investment is much larger than the earlier government commitment connected to the current AI model competition. When that competition began, the government pledged 530 billion won (about $390 million) to five selected companies. The new plan roughly multiplies that figure by nine.
The scale matters because advanced AI systems require more than research talent. They also need compute capacity, model development, testing, deployment channels and a market willing to adopt them. The proposed funding does not guarantee success, but it shows that South Korea is treating model development as a national industrial project rather than a narrow research program.
The funding would not simply extend the current contest
The current competition includes LG AI Research, SK Telecom, and Upstage. However, the two finalists in that competition would not automatically receive the new money. Instead, the government plans to open a new competition, and startups would be allowed to enter.
That detail is important because it keeps the next phase from becoming a closed follow-on award. It also suggests that the government wants to compare a wider field of possible builders before deciding where the larger pool of capital should go.
The plan also includes a second funding track. That track would support adoption of Korean-built models across the country’s economy. In practical terms, model development and model use are being treated as linked goals: South Korea wants to build stronger systems and create demand for those systems at home.
The target is China’s leading open models, not the largest US firms
Officials have acknowledged that Korea cannot compete with the largest US companies. That statement sets a clear boundary around the ambition. The goal is not to match every global AI investment plan or outspend the biggest American technology firms.
Instead, the stated competitive benchmark is the leading open models from China. That gives the plan a more focused strategic target: produce a Korean frontier model strong enough to stand against China’s best open model efforts.
The comparison also shows why the proposed amount has to be understood in context. Google, Amazon, Microsoft, and Meta are planning combined investments of around $725 billion for 2026 alone, mostly for AI data centers. South Korea’s proposed 4.7 trillion won ($3.49 billion) is large compared with its earlier government pledge, but it remains far smaller than the spending plans of the biggest US technology companies.
Compute and chips are already part of the wider picture
The proposed AI model funding is not happening in isolation. South Korean industry is already putting money into more compute capacity, which is a necessary foundation for training and running large AI systems.
In October 2025, Samsung and SK struck a deal with OpenAI to expand AI infrastructure in the country. Then in June, Samsung, SK Hynix, and the government announced a $590 billion push to scale up chip production.
Together, those moves point to a broader national effort around AI infrastructure. A homegrown frontier model needs compute, and compute depends heavily on chips and data center capacity. The proposed model investment sits on top of that industrial base rather than replacing it.
Why adoption matters as much as model building
The second funding track may be just as important as the model-building competition. A frontier model has more value if businesses, public institutions and consumers actually use systems built on it. Without adoption, even a technically strong model can struggle to become economically significant.
The source article notes that Korean consumers have been paying mostly foreign providers for AI services. In December 2025, they spent more on ChatGPT and similar subscriptions than on Netflix.
That consumer spending pattern helps explain the government’s interest in Korean-built models. If domestic users are already paying for AI tools, then South Korea has evidence of demand. The policy question is whether more of that demand can be served by local companies using Korean-developed models.
For now, the plan remains a proposal inside the 2027 budget process. If approved, it would mark a major increase in South Korea’s direct financial backing for frontier AI and a clear attempt to turn domestic model development into a broader economic strategy.