Next-Generation AI Memory Performance Optimization
Securing Products for Data Center Operations
Real-Time Validation Under Various Conditions
SK Hynix is partnering with Naver Cloud to strengthen its capabilities in developing AI (artificial intelligence) solution products. The company plans to thoroughly validate and maximize the performance of its AI-specialized products at Naver Cloud's large-scale data centers.
The two companies announced on September 10 that they had signed a memorandum of understanding (MOU) for technical cooperation at a business agreement ceremony held the previous day. Through this collaboration, SK Hynix will validate its AI-specialized product lineup-such as Compute Express Link (CXL) and Processing-In-Memory (PIM) technologies-in real time under various workload conditions using Naver Cloud’s large-scale data center infrastructure, with the goal of maximizing product performance.
Naver Cloud will apply SK Hynix’s next-generation memory solutions to its actual AI services. The aim is to increase the efficiency of GPU (graphics processing unit) utilization and reduce power consumption.
SK Hynix emphasized, "Securing products that have been validated in real data center operating environments is essential to strengthen our leadership in AI solution technologies in the global market. Through our development partnership with Naver Cloud, we will deliver AI solution products optimized for data centers and continuously discover innovative use cases that customers can experience firsthand."
With the explosive growth of generative AI services in recent times, the amount and cost of tokens processed during AI inference have increased exponentially. Tokens are the smallest units of data created by decomposing information for AI to process (learn, generate, or infer). As a result, not only are there increasing demands for memory bandwidth and capacity, but optimization between hardware-such as memory used in data centers-and software has emerged as a key factor determining the competitiveness of AI services.
SK Hynix and Naver Cloud signed a memorandum of understanding (MOU) for technical cooperation at a business agreement ceremony held on the 9th. Kim Yuwon, CEO of Naver Cloud (left), and Ahn Hyun, Head of Development at SK Hynix, are posing for a commemorative photo at the MOU signing ceremony. Provided by SK Hynix and Naver Cloud. Photo by SK Hynix and Naver Cloud.
The two companies will also pursue joint research, patent applications, and participation in international AI conferences. Through these collaborative activities, they plan to promote their technological achievements in the global market and accelerate the expansion of the AI ecosystem across the industry. Naver Cloud explained that it aims to enhance its competitiveness by acquiring optimization experience across both data center infrastructure and software, while also contributing to the strengthening of sovereign AI infrastructure based on domestic technology.
Kim Yuwon, CEO of Naver Cloud, stated, "The competitiveness of AI services is determined by optimization across the entire data center infrastructure, beyond just software. By collaborating with a leading global AI memory semiconductor company, we will strengthen our competitiveness from infrastructure to service, and provide customers with more innovative AI experiences."
Ahn Hyun, Chief Development Officer (CDO) of SK Hynix, said, "Through rigorous validation in real commercial environments, we will provide the highest level of memory solutions demanded by the global AI ecosystem, solidifying our position as a leader in AI memory. Starting with this collaboration, we will also actively expand our technology partnerships with global cloud service provider customers."
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
