Summary
ISC is already seeing AI semiconductor demand translate into reported earnings. Second-quarter 2026 revenue reached KRW 72.9 billion, while operating profit rose to KRW 21.3 billion. AI-related revenue reached KRW 59.1 billion, accounting for 81% of quarterly sales.
The company is now moving into a different stage of the cycle: production scale. ISC plans a combined KRW 100 billion investment across its Vietnam manufacturing base, Songdo operation and manufacturing-AI systems. Annual test-socket capacity is targeted to rise from roughly 2 million units in 2026 to as many as 5 million by 2029.
The opportunity is not driven by volume alone. Higher-performance GPUs, server CPUs, ASICs and premium application processors require increasingly sophisticated test sockets, while a richer high-specification product mix is supporting average selling prices.
The central investment question is therefore no longer whether ISC benefits from AI semiconductor growth. It is whether the company can scale production sharply while preserving the pricing, utilization and margins already embedded in market expectations.
Key Takeaways
- ✓ AI-related products represented 81% of ISC’s second-quarter 2026 revenue, up from 67% a year earlier. AI Compute alone generated KRW 47.7 billion.
- ✓ ISC plans to increase annual test-socket capacity from approximately 2.0 million units in 2026 to 2.7 million in 2027, 3.7 million in 2028 and as many as 5.0 million in 2029.
- ✓ ISC is estimated to hold around 90% of the global elastomer-socket market. That figure applies to the elastomer segment, not the entire global test-socket market.
- ✓ FnGuide consensus expects revenue of KRW 314.4 billion and operating profit of KRW 103.4 billion in 2026, implying a 32.9% operating margin.
- ✓ At the September 28 closing price of KRW 203,500, ISC trades at approximately 48.1x 2026E P/E and 6.9x 2026E P/B. The seven-broker consensus target price stands at KRW 220,429.
What Happened
ISC outlined a KRW 100 billion investment program aimed at expanding its ability to serve growing AI semiconductor test demand.
The program includes KRW 35 billion for its Vietnam production base, KRW 35 billion for its Songdo operation and approximately KRW 30 billion for manufacturing AI and production and quality-management systems.
Vietnam will serve as the large-volume manufacturing hub for high-performance semiconductor test sockets, including products used for AI GPUs, ASICs and CPUs. Songdo will focus more heavily on advanced product development, customized engineering, complex production and initial production stabilization.
Annual capacity is planned to increase from approximately 2 million units in 2026 to 2.7 million in 2027, 3.7 million in 2028 and as many as 5 million in 2029.
Demand is already visible in reported earnings.
Second-quarter revenue increased 41% year over year to KRW 72.9 billion, while operating profit increased 55% to KRW 21.3 billion. Operating margin reached 29%.
AI revenue reached KRW 59.1 billion. Of that, AI Compute contributed KRW 47.7 billion, AI Memory KRW 5.0 billion, On-Device AI KRW 5.4 billion and Physical AI KRW 1.0 billion.
Why It Matters
ISC’s growth opportunity is increasingly being driven by semiconductor complexity as well as semiconductor volume.
Next-generation AI accelerators combine higher processing speeds, higher power consumption and increasingly complex packages. Those characteristics raise the electrical, thermal and mechanical requirements of the interface between the semiconductor package and the test system.
The economics therefore become more attractive than a simple unit-volume story.
More AI Chips → More Complex Testing → Higher-Specification Sockets → Higher ASP
ISC is already seeing a higher mix of advanced products support average selling prices. The company is also expanding production wins beyond GPUs into server CPUs and ASICs, while entering newer testing opportunities around HBM and SOCAMM2.
Production capacity now becomes the next variable.
Technical qualification alone is not sufficient when global customers require large quantities within defined production schedules.
If ISC can combine high-end product capability with large-scale delivery, manufacturing capacity could become part of its competitive advantage rather than simply a cost of growth.
ISC’s AI thesis has already moved beyond thematic exposure.
AI accounted for 81% of second-quarter revenue, while data-center applications represented approximately 75% of sales.
The company is already generating meaningful revenue from GPU, CPU and ASIC testing rather than waiting for an emerging business to commercialize.
The next test is more demanding.
ISC has to show that substantially larger capacity can be filled without sacrificing ASP, yield or operating margin.
That distinction matters because expectations have risen with the earnings cycle.
At the September 28 reference price, ISC already trades at roughly 48x 2026 consensus earnings, while the consensus target price implies only a single-digit percentage gap from that reference price.
The AI opportunity is increasingly visible.
The next proof point is scalable economics.