The Race for AI Computing Power: Storage Chip Giants and the EUV Lithography Report

The Race for AI Computing Power: Storage Chip Giants and the EUV Lithography Report

Introduction:
The demand for high-end servers and advanced computing power has surged with the rise of generative AI. As a result, storage chip giants such as NVIDIA and AMD are vying to obtain samples of SK Hynix's next-generation HBM3E chips, which are yet to be mass-produced. This article explores the growing competition in the AI computing industry, the significance of HBM and DDR5 modules, and the key steps involved in EUV lithography.

The Battle for AI Computing Power:
NVIDIA and AMD have both requested SK Hynix to provide samples of their upcoming HBM3E chips. In their pursuit of advanced AI computing capabilities, NVIDIA has even expressed willingness to pay a premium for these samples. Currently, SK Hynix holds a 50% market share in the global HBM market, closely followed by Samsung with 40%, and Micron, a prominent US storage chip competitor, with a 10% share.

The Demand for High-Density DDR5 and HBM Products:
Since the second quarter, there has been a rapid increase in demand for high-density DDR5 and HBM products due to the emergence of generative AI. Jaejune Kim, Executive Vice President of Samsung's memory business, highlighted the strong demand for server memory in the second quarter, driven by generative AI. As capital expenditure is limited, investments in the data center sector have focused primarily on AI servers, resulting in relatively limited demand for general-purpose servers and storage.

SK Hynix's Growth in the AI Server Market:
According to Kim Woohyun, SK Hynix's sales of graphic DRAM, including HBM products, previously accounted for a single-digit percentage of the total DRAM sales. However, this proportion has rapidly grown to over 20% in the second quarter since reaching 10% in the fourth quarter of the previous year. SK Hynix has experienced significant sales growth in HBM and high-density DDR5 modules used in AI servers. The company expects HBM sales to more than double this year compared to the previous year, with further growth projected for the next year.

The Key Steps in EUV Lithography:
EUV lithography is a crucial step in manufacturing next-generation semiconductor chips. It involves generating EUV light from a high-temperature plasma produced by highly pure tin. Solid tin is melted inside a droplet generator, which continuously produces over 3 million droplets of 27µm per minute in a vacuum chamber. A 25kW carbon dioxide (CO2) laser with an average power irradiates the tin droplets in two consecutive pulses, shaping and ionizing them. Initially, thousands of watts of EUV light are generated, but due to absorption and scattering losses along the optical path, only a small fraction reaches the lithography mask. The output power and beam quality of the 13.5nm light are inferred from measurements using an indirect scintillator camera. A multi-layer reflective mirror system guides the light onto photosensitive polymers or photoresists, transferring the patterns onto silicon wafers. The reflective mirrors are protected from tin debris by a constant flow of H2 gas. An automated wafer stage positions the wafer with a resolution of ≤0.25 nm after each exposure, with 20,000 cycles per second. Overall, this process requires precise coordination among various engineering systems.

Actionable Advice:

1. Stay updated on advancements in storage chip technologies: Keep a close eye on the developments and collaborations between storage chip giants like NVIDIA, AMD, and SK Hynix to understand the latest trends in AI computing power.
2. Explore opportunities in the AI server market: With the increasing demand for AI servers, consider investing in companies that specialize in providing advanced computing solutions for generative AI applications.
3. Follow the progress of EUV lithography: As EUV lithography plays a critical role in manufacturing next-generation chips, monitor advancements in this technology to gain insights into the future capabilities and limitations of semiconductor devices.

Conclusion:
The race for AI computing power is intensifying as storage chip giants seek to acquire the latest advancements in storage chip technologies. The demand for high-density DDR5 and HBM products has surged with the rise of generative AI, with SK Hynix experiencing significant growth in the AI server market. Meanwhile, EUV lithography remains a crucial step in manufacturing next-generation semiconductor chips. By staying informed and exploring opportunities in this rapidly evolving landscape, individuals and businesses can position themselves for success in the AI computing industry.