To pack more HBM into its chips, Nvidia is accelerating the development of glass substrate technology.

Nvidia is collaborating with supply chain companies to advance glass substrate technology in order to address the growing physical limitations of AI chip packaging in terms of size, heat dissipation, and interconnectivity. If the technology achieves a breakthrough, glass substrates are expected to support larger-scale HBM integration, improving packaging density and data transmission capabilities for next-generation GPUs.

According to reports, German equipment manufacturer SCHMID disclosed at its recent first-half results presentation that it is collaborating with core companies in the supply chains of Intel, Nvidia, and AMD to jointly advance the development of glass core substrate equipment. Roland Rettenmayer, Chief Strategy Officer of SCHMID, stated that the through-glass via (TGV) metallization process still faces technical hurdles, and end-customer certification has not yet been completed.

Nvidia CEO Jensen Huang had previously engaged in extensive discussions regarding this technology approach. During his meeting with SK Group Chairman Chey Tae-won, he discussed next-generation semiconductor collaborations, including on glass substrates, and actively urged TSMC to accelerate related technology development. This indicates that Nvidia is evaluating the feasibility of incorporating glass substrates into its next-generation GPUs.

Glass substrates target AI chip packaging bottlenecks and are expected to expand HBM integration.

As AI chip performance and power consumption continue to improve, the limitations of traditional plastic substrates in terms of heat resistance and warpage control are becoming more prominent. Glass substrates, on the other hand, have higher flatness and strength, a lower coefficient of thermal expansion, and less deformation at high temperatures. They also possess good dielectric properties and low signal loss, making them a potential solution for next-generation high-performance chip packaging.

Another major advantage of glass substrates lies in their ability to scale package size. Current mainstream AI chips typically use silicon interposers to integrate the GPU and HBM side-by-side, but silicon interposers are constrained in terms of size expansion and manufacturing costs. Glass substrates hold the promise of partially or even completely replacing silicon interposers, accommodating more chips and HBMs within the same package, thereby increasing package density and shortening data transmission paths.

For AI chips, this translates to a larger HBM integration scale and higher data throughput. HBMs, which increase bandwidth by vertically stacking multiple layers of DRAM, are a crucial storage component in current AI accelerators. As model size and computing power demands continue to increase, the amount of HBM that a package can accommodate becomes a significant factor limiting AI chip performance.

Metallization of TGV remains a key factor for mass production; commercialization is not yet complete.

However, glass substrates still require breakthroughs in technology and certification before large-scale commercialization. SCHMID explicitly points out that the TGV metallization process remains one of the main technological obstacles. Since glass itself is non-conductive, the key to achieving mass production lies in how to form stable, reliable, and high-density interconnects within the substrate.

Meanwhile, the end-customer certification process is not yet complete, meaning the relevant technologies are still in the industrialization stage. SCHMID is currently developing equipment and solutions for glass substrate production and simultaneously expanding its AI server motherboard and next-generation packaging equipment business. For NVIDIA and its supply chain, whether glass substrates can ultimately enter large-scale GPU production still depends on the further maturation of manufacturing processes, reliability, and customer certification.

Risk Warning and DisclaimerInvesting involves risk; please exercise caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Any investment decisions made based on this information are at your own risk.