The race to commercialize glass substrates, seen as a game changer in next-generation semiconductor packaging, has moved beyond technical disclosures and entered a substantive mass-production readiness phase. As demand for high-performance artificial intelligence (AI) chips surges, the physical limits of existing substrates (FC-BGA) have come into focus, and glass substrates—advantageous for thermal stability and fine circuitry—have emerged as a key alternative. Market interest is shifting beyond the technology contest to the timing of mass production and the potential for revenue generation.
According to industry sources on the 10th, Intel has moved to preempt the market by presenting glass substrates as a core next-generation packaging technology. Having continued research and development (R&D) for more than 10 years, Intel invested more than $1 billion (about 1.4835 trillion won) to build R&D and pilot infrastructure at its Chandler, Arizona, facility.
The core of Intel's technology lies in combining its own advanced packaging solution, EMIB (Embedded Multi-die Interconnect Bridge), with next-generation substrate technology. EMIB is an interconnect technology that links different chips at high speed; it has been applied to some high-performance products and has demonstrated scalability.
Compared with existing materials, glass substrates enable finer wiring and large-area packaging, providing a foundation to further expand chip-to-chip interconnect structures. This makes it possible to bundle different chips—such as central processing units (CPU), graphics processing units (GPU), and high bandwidth memory (HBM)—to function like a single system, boosting data transfer speeds and reducing power loss to improve AI compute efficiency. Building on these packaging technologies and substrate innovations, Intel has presented a roadmap to integrate 1 trillion transistors per package by 2030, signaling the direction of next-generation packaging architectures.
In particular, Intel is aiming not just for a simple technological edge but to preempt the standard. If glass-substrate-based architectures take hold as the core of next-generation packaging, the designs and interconnect schemes presented by Intel are likely to harden into industry standards. This implies the de facto lead in specifications across the foundry and packaging ecosystem, a strategy interpreted as an attempt to secure a structural advantage in the AI Semiconductor market.
For now, however, the industry remains at the stage of technical implementation and reliability verification, and it expects stable, large-scale mass production to begin after 2027. While Intel is regarded as being ahead in terms of both technology and ecosystem, some note that the actual commercialization contest is only just beginning.
Samsung Electro-Mechanics is accelerating a strategy focused on customer benefit alongside the technology race. According to the industry, Samsung Electro-Mechanics recently supplied glass substrate samples to Apple following Broadcom, entering the verification pipelines of global big tech companies. It is producing prototypes through a pilot line built at the Sejong business site and, after trial production this year, is targeting mass production from 2027 onward. While Intel leads with technology direction and ecosystem, Samsung Electro-Mechanics is seen as aiming to secure commercialization results first by preempting touchpoints with key customers such as Apple.
Absolics, an SKC subsidiary, is regarded as the fastest mover in terms of production infrastructure. It has built a dedicated manufacturing base in Georgia, United States, and SKC plans to invest about 590 billion won out of roughly 1 trillion won raised through a rights offering into product development and process stabilization. However, challenges remain in obtaining global customer certifications and achieving economically viable Production yield, and the pace of commercialization appears somewhat more cautious than market expectations. Even with equipment in place, actual competitiveness is expected to hinge on process maturity and the level of yield achieved.
LG Innotek is pursuing a strategy of strengthening fundamentals while assessing market demand. It is reinforcing core technology in partnership with precision glass processor UTI, while advancing development by building a pilot production line at a domestic business site. While President Moon Hyuk-soo presented 2028 as the target for mass production, he also noted that full-fledged market demand could materialize around 2030, a stance interpreted as prioritizing technological completeness and market timing over a rushed speed race.
A semiconductor industry official said, "The glass substrate market has moved beyond research and development into a 'mass-production competition' phase, where the corporations that generate actual sales with stable Production yield will take the lead," adding, "The point at which companies pass the rigorous verification of global customers and proceed to actual supply will mark the market's true inflection point."