“In the AI era, semiconductor competitiveness depends not on who possesses more facilities, but on who can secure high-yield ‘usable production capacity’ more quickly.”

At a press conference for Semicon Korea 2026 held on February 11, Clark Tseng, Senior Director at SEMI, stated, “The demand for AI is no longer an uncertainty but a prerequisite,” emphasizing that the core competitive factor of the semiconductor industry is shifting from simple capital investment to “capabilities centered on execution speed and yield.”

Director Clark Tseng analyzed that the spread of AI is shifting the direction of semiconductor investment from a CAPEX-centric approach to actual operational production capacity and execution capabilities for equipment and materials.

The explanation is that, particularly in the memory and advanced packaging sectors, as technological difficulty increases, the focus of supply constraints is shifting from wafer input volume to packaging, testing, and customer certification schedules.

The status of the Korean semiconductor industry has also been re-examined.

South Korea's semiconductor exports reached an all-time high of $173 billion annually in 2025, driven by an increase in memory exports.

On the other hand, non-memory exports showed a stagnant trend.

Director Clark Tseng explained, “This shows that memory production capacity is being reallocated to HBM and data center storage as AI infrastructure expands.”

In particular, the fact that semiconductor exports to Taiwan increased by 65% year-on-year suggests that Korea is becoming more deeply integrated into the AI accelerator supply chain.

The expansion of investment by global cloud service providers (CSPs) is also notable.

AI-related CAPEX of the top four CSPs, including Microsoft, Google, Amazon, and Meta, is projected to reach approximately $400 billion in 2025 and expand to $650 billion in 2026.

Director Clark Tseng stated, “Both global semiconductor sales and AI-related investments will surpass $1 trillion by 2027,” emphasizing that investment in AI infrastructure is a long-term structural change rather than a short-term trend.

Amidst these trends, AI and High Performance Computing (HPC) are also driving the growth of the semiconductor equipment market.

Investment in AI and HPC-related equipment accounted for about 40% of the total in 2024, and is projected to exceed half after 2028.

In particular, investment in test equipment and advanced packaging equipment is rapidly expanding due to the increasing demand for AI accelerators and HBMs.

However, Director Clark Tseng identified power and cooling infrastructure, advanced packaging processes, and testing and burn-in capabilities as bottlenecks in the expansion of AI infrastructure.

Finally, he said, “Korea possesses strengths in the memory and advanced packaging ecosystems, and these execution capabilities will determine the speed of global AI infrastructure expansion.” "..." he said, predicting that the future growth of the Korean semiconductor industry will shift beyond simple volume expansion to integrated competitiveness at the system level.