This page was machine-translated and may differ from the original. View original
Kwak No-jeong, CEO of SK Hynix, said, "As a 'full-stack AI memory creator,' we will deliver value that exceeds customer expectations."
▲SK Hynix CEO Kwak No-jeong is giving a keynote speech on the topic of “SK Hynix’s new vision and technology in the AI era” at the “SK AI Summit 2025.”
Full-stack AI memory lineup expands NAND stacking bandwidth and implements high-capacity storage.
Expanding the AI memory ecosystem through collaborations with big tech companies like NVIDIA, OpenAI, and TSMC.
Expanding the AI memory ecosystem through collaborations with big tech companies like NVIDIA, OpenAI, and TSMC.
As a full-stack AI memory creator, we will work with our customers to solve their pressing challenges and deliver value beyond what they desire.
At the 'SK AI Summit 2025' held on the 3rd, SK Hynix President and CEO Kwak No-jung announced a new vision to lead the AI era, 'Full Stack AI Memory Creator,' and presented the future direction of the global memory market.
CEO Kwak No-jeong emphasized, “Until now, SK Hynix has played the role of a ‘full-stack AI memory provider’ that supplies products that customers want in a timely manner,” and “Going forward, we will evolve into a creator that creates AI memory as a co-architect and ecosystem contributor.”
In the AI era, memory semiconductors are important.As the number of nodes grows, the 'memory wall' problem, where memory performance becomes a bottleneck along with the increase in data movement, is emerging. Accordingly, memory is positioned as a core value product in the AI industry, rather than a simple component.
CEO Kwak No-jeong unveiled SK hynix's full-stack AI memory lineup to address AI inference bottlenecks and efficiently utilize computing resources.
Custom HBM maximizes GPU and ASIC performance and improves power efficiency with customer-specific designs.
AI DRAM (AI-D) is being developed in three directions: optimization (AI-D O), breakthrough (AI-D B), and expansion (AI-D E).
AI NAND (AI-N) is a storage solution focused on performance (P), bandwidth (B), and high density (D), and is targeting sample launch in late 2026.
In particular, it includes various technologies such as MRDIMM, SOCAMM2, LPDDR5R, CXL-based CMM, and PIM, and is also pursuing bandwidth expansion through NAND stacking and implementation of petabyte-level high-capacity storage.
CEO Kwak No-jeong said, “SK Hynix is not only collaborating with NVIDIA on HBM, but is also pursuing AI manufacturing innovation using Omniverse and Digital Twin.” He added, “We are working on a long-term collaboration with OpenAI to supply high-performance memory, and we are collaborating with TSMC on the development of next-generation HBM base dies.”
He also mentioned, “We are accelerating the expansion of the AI memory ecosystem by collaborating with SanDisk for international standardization of HBF (High Bandwidth Flash) and Naver Cloud for data center optimization.”
CEO Kwak No-jeong said, “In the AI era, companies that create greater synergy through collaboration with customers and partners will succeed.”o;SK Hynix stated, “We will prioritize customer satisfaction and strengthen technological advancement and ecosystem cooperation to pioneer the future together.”
본 기사에 대한 정정·반론·추후보도 청구는 보도 청구 안내를, 그간 게재된 보도문은 정정·반론보도 모아보기를 참고해 주세요.














