Insights

As the US and China engage in all-out competition for AI technology supremacy, an urgent analysis of Korea's positioning in the AI industry has become unavoidable.

Samjong KPMG Economic Research Institute's Issue Monitor compares US and Chinese AI competitiveness across six dimensions — investment ecosystem, government support, AI infrastructure, talent and IP assets, mid-term applications, and long-term competitive arenas — and draws out the opportunities and risks for Korea.

Performance gap nearly closes in a single year. In 2023, the US held a sizeable lead on key AI performance benchmarks. The gap in language understanding (MMLU) accuracy was 17.5 percentage points and in coding (HumanEval) pass rate was 31.6 percentage points. By 2024, the MMLU gap had collapsed to 0.3 percentage points and HumanEval to 3.7. In just one year, Chinese AI models reached near-parity with the US across many core areas. This catch-up has intensified concern in US politics and triggered additional export controls and technology restrictions.

Six dimensions: US ahead on capital and infrastructure, China racing on application and scale. In the investment ecosystem, the US retains the lead with OpenAI, Anthropic, Google DeepMind, and others. But China's machine-learning market is projected to grow roughly tenfold — from USD 9.72 billion in 2024 to USD 96.17 billion by 2030. Longer-term, the defense AI race is heating up as China aims to intelligentize the PLA by 2035 and embed AI as a core combat capability.

Korea's AI ecosystem: KRW 969.4 billion in 2024 — a record high, with strategic leap-frogging as the key. Korea's AI startup VC investment grew from KRW 343.5 billion in 2020 to KRW 969.4 billion in 2024 — a record high, up 22.1% year-on-year. Semiconductor-based players such as FuriosaAI and DeepX, and LLM platform companies like Wrtn Technologies and Upstage, secured large funding rounds. The report argues that an “industry-specialized AI strategy” — embedding AI into Korea's strong industries (semiconductors, manufacturing, autos) to enable process automation and quality prediction — can be Korea's differentiating path within global competition.

What this means for M&A. The deeper the US-China AI competition runs, the more distinctive M&A opportunities open for Korean firms. First, with rising technology nationalism on both sides, Korean AI companies' attractiveness as “neutral and trustworthy” cross-border M&A partners is set to rise. Second, demand for AI internalization in Korea's flagship industries (semiconductors, manufacturing, biotech) will accelerate strategic acquisitions of domain-specialized AI solution startups. Third, as global AI infrastructure competition lifts valuations of data center, power, and cooling infrastructure firms, related infrastructure companies will emerge as new deal-sourcing targets.