🔗 LG Innotek: The Gap Between Its AI Substrate Roadmap and Mass-Production Timeline - 2026.09.10

Hi, it’s Dotori. Today, the key issue for LG Innotek is not simply the fact that it showcased substrates for AI semiconductors, but how the disclosed technologies will lead to actual mass-production schedules. At KPCA Show 2026, large-area, high-layer-count substrates and products for AI accelerators took center stage, but the exhibition floor is only the starting line; the real weight of the business lies in what comes next.

🔗 Let’s start with the big picture

LG Innotek presented its semiconductor substrate and packaging business strategy alongside Samsung Electro-Mechanics by unveiling a roadmap related to semiconductor packaging ^[3]. At this exhibition, LG Innotek unveiled an FC-BGA for AI accelerators with a side length exceeding 100mm, introducing it as a product needed for high-performance GPUs and NPUs ^[1].

This connection is the key. As AI computing grows, we have to look not only at accelerator performance but also at the area, layer count, and interconnection structure of the package substrate supporting it. It is less like enlarging just one gear and more like a system that works only when the surrounding gears mesh together as well.

⚙️ 2027 and 2028: Different Validation Phases

According to the materials currently available, LG Innotek plans to begin mass production of FC-BGA for AI semiconductors in 2027, while targeting mass production of next-generation glass substrates in 2028 ^[2]. The important point here is not to lump the two technologies together. FC-BGA is a product for which an earlier mass-production schedule has been presented, while glass substrates belong to a next-generation area with a later target date.

That means the questions the market will verify are also different. For FC-BGA, the key issues are whether mass production actually begins on schedule and whether the supply range expands enough to meet demand for high-performance GPUs and NPUs. For glass substrates, the remaining variables are whether the 2028 target becomes a concrete business schedule and whether the technology displayed moves into the production stage. However impressive the technology names may sound, factories still bring the practical tests of yield and customer certification.

🏭 Which Part of the Value Chain Will Appear First?

If we divide differentiated benefits across the broader industry landscape, they can be summarized as 1st priority: FC-BGA for AI accelerators, 2nd priority: large-area, high-layer-count package substrates, and 3rd priority: next-generation glass substrates. This distinction reflects the specificity of the publicly disclosed products and schedules; it does not mean business results are assured.

KPCA Show 2026 was held at Songdo Convensia in Incheon from September 9 to 11, 2026, where LG Innotek and Samsung Electro-Mechanics showcased large-area, high-layer-count substrates and next-generation package-substrate technologies responding to demand for AI semiconductors ^[4]. The fact that the technologies were unveiled on the same stage as competitors is also meaningful. The basis for comparison is now shifting away from who drew the grander blueprint and toward who can connect mass production with expanded supply more concretely.

📚 The Gap Dotori Sees Ahead

Based on these materials alone, Dotori believes LG Innotek’s central question is not the disclosure of technology but the connection between its mass-production schedule and actual supply expansion. With targets set for FC-BGA mass production in 2027 and glass-substrate mass production in 2028, confirmation of customer adoption and production progress will inevitably become more important in the next phase.

However, these materials alone do not reveal order volumes by product, profitability, or actual customers. We therefore need to separate the broad trend of AI substrates from each company’s execution speed. The exhibition lights were bright, and now the question is whether that light will extend all the way to the production line. How the 2027 mass-production target becomes more concrete, and what additional information will support the 2028 glass-substrate target, remain open variables.

📰 Reference News (4)

1

[Business Today] Samsung Electronics Partners with Mistral AI on Semiconductor AI, and More - DealSite Economy TV

네이버뉴스 · 2026.09.09

📰 View original

2

LG Innotek, Too

네이버뉴스 · 2026.09.10

📰 View original

3

[On Site] “Changing the Semiconductor Landscape”… Samsung Electro-Mechanics and LG Innotek Unveil Their ‘Packaging Roadmap’ - Financial Post

파이낸셜포스트 · 2026.09.10

📰 View original

4

[Semiconductor Radar] Samsung Electro-Mechanics and LG Innotek Put AI Accelerator and Glass-Substrate Technologies Front and Center at ‘KPCA Show’

네이버뉴스 · 2026.09.10

📰 View original