Korea 2027 R&D Budget: 39.5 Trillion Won, 4.1 Trillion for AI

Korea’s 2027 government R&D budget proposal rises 11.2%, with 4.1 trillion won for AI and 9.5 trillion won for three “mega projects” and seven “SEED” fields. Basic research grows 5.9%. A shorter review process for large programs and conflicts of interest in project vetting are now the key issues in parliamentary review.

According to the proposal released by the Ministry of Science and ICT on Sept. 1, 2026, major R&D spending allocated and coordinated by the ministry’s Office of Science, Technology and Innovation comes to 33.8 trillion won, up 11.5% from the previous year. General R&D drawn up by the Ministry of Planning and Budget adds 5.7 trillion won. The government plans a record 3.3 trillion won for new programs and large-scale R&D centered on national strategic technologies. The 39.5 trillion won total is still a government proposal that must pass the National Assembly, and the size of the budget should be judged separately from actual spending and research outcomes.

4.1 trillion won for AI R&D: money concentrated on strategic technology

The biggest change is the expansion of AI R&D, budgeted at 4.1 trillion won. That is 1.7 trillion won more than roughly 2.4 trillion won in 2026, an increase of 69.1%. The plan is to raise the competitiveness of Korea’s AI ecosystem by investing in world models trained on large-scale synthetic data, general-purpose robot foundation models and core technology for AI data centers (AIDC).

AI investment is not limited to developing software models. It also requires semiconductors for AI computing, power, cooling and networking technology for data centers, sensors and actuators for robots, and technology to collect and use data on real industrial sites. Physical AI in particular connects AI trained in digital environments to physical systems such as robots and manufacturing equipment. The government has set targets to strengthen physical AI competitiveness by 2030 and to raise the share of domestically produced core robot components.

A bigger AI budget does not by itself raise Korea’s technological competitiveness at the same rate. What matters is how efficiently domestic researchers use high-performance computing once it is secured, whether home-grown models and training data accumulate, and whether companies and research institutes solve real industrial problems. GPUs and data centers are the foundation of R&D, not its final output.

In AI data centers and semiconductors, the balance between importing foreign equipment and developing domestic technology also deserves attention. If dependence on outside core equipment and software continues, larger investment may do little for technological self-reliance. If, on the other hand, Korean companies, universities and government-funded research institutes build experience with the equipment along with design, software and systems-integration capabilities, infrastructure spending could translate into long-term industrial competitiveness.

9.5 trillion won for three mega projects and seven SEED fields

The government presented the three mega projects and seven SEED fields as the backbone of its strategic technology investment. The 2027 proposal allocates 5.8 trillion won to the mega projects and 3.7 trillion won to the SEED fields, 9.5 trillion won in total.

Area	2027 proposal	Main goals
Three mega projects	5.8 trillion won	Strategic technologies including AI, semiconductors, physical AI and next-generation security and networks
Seven SEED fields	3.7 trillion won	Future energy and nuclear, advanced bio, aerospace and marine, quantum, advanced supply chains and more
AI R&D	4.1 trillion won	World models, robot foundation models, AI data center technology
Regional R&D	4.5 trillion won	Region-led R&D and technology transition for regional industry
Basic research and universities	3.6 trillion won	Strengthening the basic research base and university research ecosystem
Defense and K-defense	5.0 trillion won	Defense technology and defense industry competitiveness
Young researchers and talent	1.8 trillion won	Support for early-career researchers and science and engineering talent

Source: Ministry of Science and ICT, 2027 government R&D budget proposal.

Among the SEED fields, the proposal allocates 1.7 trillion won to future energy and nuclear, 2.2 trillion won to advanced bio, 2 trillion won to aerospace and marine, and 370 billion won to quantum. Aerospace and marine spending rises 35.4% from the previous year. The government plans to expand support for fields that need long-term technology accumulation, such as a lunar lander, eco-friendly icebreakers, next-generation reactors and quantum technology.

Concentrated investment in strategic technology can pool national research capacity and strengthen industrial competitiveness. But in a budget built around large projects, it is also necessary to ask how priorities were set, how the roles of existing and new programs differ, and whether technology is transferred to the private sector when programs end. The more money is concentrated, the more specific performance evaluation needs to be.

Shorter reviews for large R&D programs: balancing speed and scrutiny

The government proposed cutting the review period for large new R&D programs with total budgets of 100 billion won or more from about two years to around three months. It said 23 programs worth about 1.1 trillion won in the 2027 proposal reflect the shortened procedure.

Shorter reviews reduce the risk of missing the right moment to invest in fast-moving fields. In areas with short technology cycles, such as AI and semiconductors, critics have long argued that years of administrative procedure undermine timeliness.

Shorter reviews must not, however, weaken the assessment of feasibility and risk. Large R&D programs have to account not only for upfront costs but also for operating costs, staffing, demonstration costs and the cost of revising plans as technology changes. Even if programs start sooner, it must still be possible to verify technical feasibility, duplicate spending and the division of roles with private investment.

With new-program spending rising to 3.3 trillion won, selection criteria and evaluation results should be published, and follow-on funding should be adjusted according to stage-by-stage results. Starting quickly and managing the risk of failure are not conflicting goals. Clear criteria set in advance and institutionalized interim reviews allow fast investment and responsible budgeting at the same time.

3.6 trillion won for basic research: the total versus what labs feel

The government budgeted 3.6 trillion won for basic research, a record and 5.9% more than the previous year. It also plans to more than double funding for grassroots basic research, steadily increase the number of grants and support early-career researchers.

In growth terms, however, basic research is rising more slowly than total R&D (11.2%), and far more slowly than AI (69.1%). That gap alone does not prove basic research is being neglected. Because basic research pays off over long periods, single-year growth should be read alongside the number of grants, funding per grant, grant duration and researchers’ job stability.

Basic research explores questions that are hard to commercialize today and widens future technological options. Work with no direct link to current industrial demand can become the starting point of new technologies and industries. Evaluations of basic research support should therefore weigh originality, potential for follow-on research and whether early-career researchers can work independently, not just short-term patents or revenue.

Researchers’ working and living conditions also matter. Without enough research time and stable support for graduate students and postdoctoral researchers, more funding alone will not raise research capacity. The state of equipment, laboratories, dormitories and other facilities also affects whether research can be sustained. Claims that a specific facility received no funding, however, should be checked against the institution’s budget documents and program-level allocations.

Planners who also review: how effective are conflict-of-interest rules?

Alongside the size of the budget, the fairness of how R&D programs are planned and evaluated also needs scrutiny. At the National Assembly Science, ICT, Broadcasting and Communications Committee’s audit on Oct. 6, 2026, lawmakers raised a conflict-of-interest problem: experts who helped plan national R&D programs could serve as expert reviewers of those same programs’ budgets.

According to news reports, Democratic Party lawmaker Lee Yeon-hee compared 490 expert reviewers active over the past six years with 281 planning reports for new R&D programs, and found 38 people who had helped plan a total of 41 programs and then served as reviewers of the related budgets. A next-generation network (6G) industrial technology program with a total budget of 440.7 billion won was cited as one such case.

The central point is that people who helped design a program may face a conflict of interest when they then review its necessity and budget.

Even with rules for disqualification, challenge and recusal in place, their effectiveness needs to be examined if they are not actually invoked. Further checks are needed on the relevant reviewers’ actual participation, the minutes of review meetings and how disqualification criteria are applied.

At the audit, Minister of Science and ICT Bae Kyung-hoon acknowledged a potential conflict of interest when experts who planned a program review its budget, and indicated the ministry would prepare improvements. Those should include mandatory disclosure of interests, checks of planning history before review, and specific criteria for excluding the reviewers concerned from deliberation and decisions on related items. Where expertise is needed as an exception, safeguards such as independent external review and published minutes could be added.

Three issues for parliamentary budget review

The proposal now moves to the National Assembly’s Special Committee on Budget and Accounts and the standing committees. The review schedule for the 2027 budget, finalized by the ruling and opposition parties on Oct. 23, runs for about a month starting with a public hearing on Oct. 30. For a record 39.5 trillion won to be used well in research, three questions need answers first.

First, whether AI research spending goes beyond buying hardware. Much of the roughly 1.7 trillion won increase in AI R&D is concentrated on securing high-performance GPUs and building infrastructure such as AI data centers. Careful design review is needed to ensure the money leads to home-grown foundation models and real technology accumulation in manufacturing and robotics, not just imported equipment and construction.

Second, basic research growth at half the overall rate, and how labs experience it. The government stresses that basic research reached a record 3.6 trillion won, but its 5.9% growth is about half the 11.2% rise in total R&D. The audit also revealed that while enrollment quotas were raised at Korea’s four science and technology institutes, including UNIST and DGIST, next year’s proposal included no funding to improve the aging dormitories where master’s and doctoral researchers live. Concerns that basic capacity and research conditions are being overshadowed by headline figures for mega projects need to be addressed.

Third, transparent allocation of 3.3 trillion won in new programs. Next year’s new-program budget alone reaches a record 3.3 trillion won, and preliminary feasibility reviews for large programs have been cut to three months. With faster reviews and large sums released, enforceable conflict-of-interest safeguards that rule out the “planners reviewing their own programs” problem raised in the audit must be in place before the budget is finalized if the allocation is to be trusted.

Upcoming schedule and review items

Date	Body	Schedule and review items
Oct. 28 (Wed) – 30 (Fri)	Standing committees	End of the parliamentary audit, with audits by committees such as House Steering, Intelligence and Gender Equality and Family
Oct. 30 (Fri), 2 p.m.	Budget committee	Public hearing on the 2027 budget and fund management plans; expert testimony and start of full review
First to mid-November	Standing committees	Preliminary review by committee, including the Science, ICT, Broadcasting and Communications Committee
Nov. 4 (Wed) – 5 (Thu)	Budget committee	General policy questioning of the prime minister, deputy prime minister for the economy and other ministers
Nov. 9 (Mon) – 10 (Tue)	Budget committee	Review of economic ministries, including planning and budget, finance and economy, and trade, industry and energy
Nov. 11 (Wed) – 12 (Thu)	Budget committee	Review of non-economic ministries, including science and ICT, education and defense
Nov. 16 (Mon) – end of November	Budget committee	Budget adjustment subcommittee: final line-by-line cuts and increases, then full committee vote
Dec. 2 (Wed)	Plenary session	Legal deadline for passing the budget