When people talk about AI chips, they usually focus on the processor itself.
They talk about performance, memory, data centres, power use and the companies designing the chips.
But as artificial intelligence chips become larger and more complex, another part of the semiconductor structure is becoming more important: the package substrate.
A package substrate is the base layer that connects a chip to the rest of the system.
It helps carry electrical signals, deliver power and keep the chip package physically stable.
Without a reliable substrate, even the most advanced chip cannot perform properly.
This is why glass substrates are attracting attention in advanced semiconductor packaging.
They are not yet a mainstream replacement for organic substrates.
They are better understood as a next-generation candidate material being developed for larger, denser and more demanding AI chip packages.
From a materials engineering point of view, the idea is simple.
When the chip grows larger, the material beneath it must become more precise.
Packaging Is No Longer the Final Detail
AI accelerators are different from ordinary chips.
They often combine logic chips, high-bandwidth memory and multiple chiplets inside one package.
As these packages become larger, the substrate must handle more electrical connections, more power delivery demands and more physical stress.
One important challenge is warpage.
Warpage means that a substrate bends or changes shape during heating, cooling or assembly.
In advanced semiconductor packaging, even a small deformation can affect the alignment of fine electrical connections.
This can make manufacturing more difficult and may affect long-term reliability.
For many years, packaging was often treated as the later stage of semiconductor production.
That view is becoming outdated.
The chip itself still matters.
But how that chip is connected, powered and supported now matters much more than before.
In the AI era, packaging is not just the box around the technology.
It is part of the technology.
What Makes Glass Substrates Interesting
Glass substrates are being developed because they offer several useful properties for advanced packaging.
First, glass can provide a very flat surface.
This helps engineers create fine wiring patterns.
Second, glass has strong dimensional stability.
This is important when chip packages become larger and require accurate alignment.
Third, glass can support through-glass vias, also known as TGVs.
These are tiny vertical paths that allow electrical signals to pass through the glass layer.
Fourth, glass may help improve signal transmission in high-performance computing systems.
This matters for AI servers, where speed, power efficiency and stable communication between components are important.
Intel has described glass substrates as a promising material for future advanced packaging. The company has said that glass substrates can tolerate higher temperatures, reduce pattern distortion and support much higher interconnect density compared with some existing substrate materials.
However, this point needs caution.
A material can look strong in a technical roadmap and still face many years of manufacturing work.
Glass substrates are promising.
They are not yet used everywhere in commercial AI chips.
Why Korea Is Connected to This Story
South Korea is closely connected to the glass substrate discussion because it already has strong semiconductor, display, materials and precision manufacturing industries.
One company connected to this field is Absolics, a subsidiary of Korea-based SKC.
Absolics has been developing glass substrates for advanced semiconductor packaging and operates a facility in Covington, Georgia.
The U.S. National Institute of Standards and Technology has described federal support for Absolics as part of the CHIPS for America effort to strengthen advanced packaging supply chains.
This matters because glass substrates are not only a Korean industry story.
They are also part of a wider attempt to build more resilient semiconductor supply chains outside the most concentrated production hubs.
Samsung Electro-Mechanics is another Korean company moving into this field.
In 2025, the company announced cooperation with Japan’s Sumitomo Chemical Group to establish a joint venture for glass core materials.
A glass core is one of the key parts of a glass substrate used in package substrates.
These moves do not mean Korea already controls the market.
They show that Korean companies are trying to secure an early position in a field that could become more important as AI chips continue to evolve.
That distinction matters.
Early investment is not the same as guaranteed leadership.
Why This Topic Matters Outside Korea
For readers outside Korea, glass substrates matter for three reasons.
First, they show that the AI boom is changing the entire semiconductor supply chain, not only chip design.
Second, they show Korea’s attempt to expand beyond memory chips into advanced packaging materials.
Third, they help explain why the next stage of semiconductor competition may depend on materials, substrates and packaging technology as much as on transistor size.
In simple terms, future AI performance will not depend only on the chip itself.
It will also depend on how well the chip is connected, powered and supported.
The public may see the processor name.
Engineers must look at the full stack.
That stack includes the material below the chip.
The Challenges Are Still Significant
Glass substrates are promising, but several barriers remain.
The first is cost.
Organic substrates are already widely used and supported by mature manufacturing lines. Glass substrates require new equipment, new processes and stable yield control.
The second challenge is manufacturing precision.
Creating through-glass vias and forming reliable electrical connections inside glass requires advanced processing technology.
The third challenge is customer adoption.
Major chipmakers need to test reliability, performance, supply stability and cost before using a new substrate material in commercial products.
The fourth challenge is scale.
A material can look promising in development, but mass production is a different test.
Yield, defect control and long-term reliability will decide whether glass substrates can move from pilot lines to broader commercial use.
For this reason, glass substrates should be understood as a next-generation candidate material.
They should not be described as an immediate replacement for all existing substrates.
Sustainability Needs Careful Language
Sustainability should be discussed carefully in this topic.
Glass may offer some long-term environmental advantages if manufacturing processes become cleaner and recycling systems improve.
But it would be too early to describe glass substrates as automatically sustainable without full lifecycle evidence.
A material is not sustainable only because it sounds cleaner.
Manufacturing energy, chemical use, yield loss, recycling systems and supply-chain practices all matter.
Sustainability regulations are also changing.
EU sustainability reporting and due diligence rules have gone through postponements and simplification measures, so they should not be presented as a direct 2026 reason for glass substrate adoption.
A safer way to explain the issue is this: as global technology companies face more pressure to manage supply-chain risks and environmental reporting, materials and manufacturing methods may become more important in procurement decisions.
That does not mean sustainability is the main reason glass substrates are being developed.
The main reason is performance.
But environmental reporting and supply-chain transparency may still influence how future materials are evaluated.
What to Watch Next
Several developments will decide how important glass substrates become.
Will pilot lines move into stable mass production?
Will major chipmakers confirm commercial adoption?
Will production costs fall?
Will glass substrates prove reliable in large AI packages?
Will Korean suppliers secure long-term customers?
Will organic substrates remain more cost-effective for most applications?
Will TGV processing reach the yield levels needed for commercial scale?
These questions matter more than short-term hype.
In materials engineering, the attractive property is only the beginning.
The real test is whether that property can survive cost, yield, heat, stress, handling and customer qualification.
Final Takeaway
Glass substrates are becoming an important topic in advanced semiconductor packaging because AI chips are becoming larger, more complex and more demanding.
They offer useful properties such as flatness, dimensional stability and the potential for dense interconnections.
At the same time, they still face cost, manufacturing and adoption challenges.
For South Korea, this is a meaningful semiconductor story.
It shows how Korean companies are trying to strengthen their role in the next stage of the AI chip supply chain.
But this should be read as an early-stage materials and packaging story, not as a settled market outcome.
The chip itself may receive most of the public attention.
But in the coming years, the material underneath the chip could become one of the quiet foundations of AI hardware.
Technology and Industry Information Notice
This article is for general technology, materials and semiconductor industry information only.
It does not provide investment advice, stock analysis, company recommendations, procurement guidance or environmental certification.
Glass substrate development, advanced packaging adoption, company plans, CHIPS Act funding, EU sustainability rules, manufacturing yields and commercial timelines can change.
Readers should check official company disclosures, government notices, technical publications and qualified professional advice before making business, investment or policy decisions.