SK hynix is showcasing its cutting-edge AI memory technologies at GPU Technology Conference (GTC) 2025 held in San Jose from March 17–21.

Hosted annually by NVIDIA, GTC is one of the world’s leading AI conferences. This year’s event, held under the theme “What’s Next in AI Starts Here,” features over 1,000 sessions and 300 exhibitions highlighting the latest advancements and solutions in AI technology.

Visitors gather to see the latest AI memory innovations at SK hynix’s booth

At the conference, SK hynix is presenting a booth under the theme “MEMORY, POWERING AI and TOMORROW,” showcasing memory products that accelerate AI innovation. These products are displayed in four sections: AI/Data Centers (DC), HBM¹, On-Device, and Automotive. In particular, the HBM section debuted a model of the 12-layer HBM4, which is currently under development, demonstrating the company’s overwhelming technological leadership.

¹High Bandwidth Memory (HBM): A high-value, high-performance product that significantly enhances data processing speeds by vertically stacking multiple DRAM chips, surpassing the capabilities of traditional DRAM. HBM has evolved through six generations, starting with the original HBM and followed by HBM2, HBM2E, HBM3, HBM3E, and HBM4.

The AI/Data Centers section features high-performance eSSDs, DIMM solutions, and more

The AI/DC section features enterprise SSDs (eSSDs) optimized for large-scale data processing and storage, including PEB110, PS1010, and PEB9010, alongside data center DIMM² products such as DDR5 RDIMM and MRDIMM. The world’s highest-specification graphics memory, GDDR7³, is also on display. Additionally, next-generation AI memory products are drawing attention, including CMM-DDR5⁴, a DDR5-based CXL⁵ product, and SOCAMM⁶, a low-power DRAM module for AI servers. The company plans to begin mass production of SOCAMM in line with the market’s emergence, leveraging its strengths in LPDDR5X⁷ and server memory to expand its AI memory portfolio.

²Dual In-line Memory Module (DIMM): A module in which multiple DRAM chips are combined on a substrate.
³Graphics DDR (GDDR): A standard specification of graphics DRAM defined by JEDEC and specialized for processing graphics more quickly. It is now one of the most popular memory chips for AI and big data applications.
⁴CXL Memory Module-DDR5 (CMM-DDR5): A next-generation DDR5 memory module utilizing CXL technology to boost bandwidth and performance for AI, cloud, and high-performance computing.
⁵Compute Express Link (CXL): A next-generation interface that efficiently connects CPUs/GPUs and memory in high-performance computing systems, supporting large-scale, ultra-fast computations. Applying CXL to existing memory modules can expand capacity by more than 10 times.
⁶Small Outline Compression Attached Memory Module (SOCAMM): A next-generation memory module based on LPDDR5X, featuring a smaller form factor than traditional server memory modules while offering higher power efficiency.
⁷Low Power Double Data Rate 5X (LPDDR5X): Low-power DRAM for mobile devices, including smartphones and tablets, aimed at minimizing power consumption and featuring low voltage operation. LPDDR5X is the seventh-generation LPDDR product.

The company unveiled a model of its HBM4 at GTC 2025

The highlight of the HBM section is undoubtedly HBM4. Unveiled for the first time in the world at GTC 2025, the in-development HBM4 is set to offer significantly enhanced base die performance and vastly reduced power consumption. Following the introduction of this latest AI memory solution, HBM4, SK hynix will further strengthen its industry leadership. Additionally, the company is showcasing the industry’s best-performing and highest-capacity 12-layer HBM3E available today, alongside NVIDIA’s GB200 Grace^TM Blackwell Superchip.

The On-Device section features various products tailored for on-device and AI applications, including LPDDR6 and LPCAMM2

Visitors to the On-Device section can explore the company’s high-performance, low-power memory solutions designed for on-device data processing and AI computing. These include LPDDR6, LPCAMM2⁸, PCB01, and ZUFS⁹ 4.0.

⁸Low Power Compression Attached Memory Module 2 (LPCAMM2): LPDDR5X-based module solution that offers power efficiency and high performance as well as space savings. It offers performance levels equivalent to two DDR5 SODIMMs, making it optimized for on-device AI.
⁹Zoned Universal Flash Storage (ZUFS): An advanced version of UFS that improves data management efficiency by organizing data into zones, optimizing transfer between the operating system and storage device.

Various vehicle memory solutions are also on display in the Automotive section

Meanwhile, at the Automotive section, SK hynix is presenting a range of vehicle memory solutions aimed at advancing the era of autonomous mobility. These include HBM2E, LPDDR5, UFS¹⁰ 3.1, and Gen4 SSD tailored for the automotive industry.

¹⁰Universal Flash Storage (UFS): A breakthrough type of flash memory that can simultaneously read and write data. Due to its low-power consumption, high-performance and reliability, UFS is widely applied in mobile devices.

As well as its exhibition at GTC 2025, SK hynix has also prepared a virtual booth. Online visitors can browse a web brochure detailing the company’s achievements in 2024 and information on key products including server solutions, the high-capacity SSD PS1012, GDDR7, and CMM-DDR5. Furthermore, there are also videos of innovative products such as HBM, 1c DDR5¹¹, and automotive memory, emphasizing SK hynix’s technological leadership.

¹¹1c DDR5: A 16 gigabit DDR5 DRAM built using the 1c node, the sixth generation of the 10 nm DRAM process. The 10 nm DRAM process was developed in the order of 1x-1y-1z-1a-1b-1c. In October 2024, SK hynix became the first in the world to successfully develop the 1c DDR5 DRAM.

Jungsoo Park of HBM Product Planning presenting on the importance of HBM in the AI era

The company also shared its insights with industry professionals during the event’s presentation sessions. Technical Leader Jungsoo Park of HBM Product Planning presented on “HBM: Backbone of High-Performance Computing and AI,” highlighting the evolution of HBM technology and SK hynix’s unrivaled leadership in the field.

Kihong Kim of Mobility Business giving a talk on SK hynix’s automotive memory

Meanwhile, in an online session titled “Preparing for the Future: Automotive Memory and Storage Requirements”, Technical Leader Kihong Kim of Mobility Business introduced the characteristics of memory and storage in the automotive industry and SK hynix’s related strategic approach.

As AI technology continues its rapid evolution, the role of memory is becoming increasingly vital. At GTC 2025, SK hynix is underlining its AI memory leadership by introducing solutions optimized for the AI era. Looking ahead, the company will further strengthen collaboration with global AI companies to remain at the forefront of AI memory innovation.

The post SK hynix Showcases Unrivaled AI Memory Leadership at GTC 2025 first appeared on SK hynix Newsroom.

Figure 1. SK hynix’s exhibition booth at the conference

Showcasing the Industry’s Highest Standard of AI Memory

The AI revolution has continued to pick up pace as AI technologies spread their reach into various industries. In response, SK hynix is developing AI memory solutions capable of handling the vast amounts of data and processing power required by AI. At GTC 2024, the company is displaying some of these products, including its 12-layer HBM3E and Compute Express Link (CXL)¹, under the slogan “Memory, The Power of AI”.

¹Compute Express Link (CXL): A PCIe-based next-generation interconnect protocol on which high-performance computing systems are based.

Figure 2. HBM3E, the highest-specification DRAM for AI applications

HBM3E, the fifth generation of HBM², is the highest-specification DRAM for AI applications on the market. It offers the industry’s highest capacity of 36 gigabytes (GB), a processing speed of 1.18 terabytes (TB) per second, and exceptional heat dissipation, making it particularly suitable for AI systems. On March 19, SK hynix announced it had become the first in the industry to mass-produce HBM3E.

²High Bandwidth Memory (HBM): A high-value, high-performance product that possesses much higher data processing speeds compared to existing DRAMs by vertically connecting multiple DRAMs with through-silicon via (TSV).

Figure 3. CXL and GDDR7 are set to become key components of AI systems

Displayed alongside HBM3E is CXL, a standardized interface that increases the efficiency of utilizing CPUs, GPUs, accelerators, and memory. CXL is renowned for its ability to expand memory and improve the performance of AI applications. It is therefore set to play a key role in the AI era along with HBM3E. Visitors to SK hynix’s booth can also check out some never-before-seen products, including the next-generation GDDR7 DRAM specialized for video and graphics processing. Compared to its predecessor GDDR6, the latest GDDR³ product offers double the maximum bandwidth (128 GB per second), a 40% improvement in power efficiency, and enhanced visuals due to its expanded memory density. Set to be used in graphics cards and GPUs, GDDR7 will be integral for AI and advanced technologies which require rapid and large-capacity data processing.

³Graphics DDR (GDDR): A standard specification of graphics DRAM defined by the Joint Electron Device Engineering Council (JEDEC) and specialized for processing graphics more quickly. Its generation shifted from GDDR3, GDDR5 and GDDR5X to GDDR6. It’s now one of the most popular memory chips for AI and big data applications.

Figure 4. A PCB01-based consumer SSD customized for AI

SK hynix also revealed a new consumer product based on its industry-leading Gen 5 solid-state drive (SSD), PCB01. SK hynix plans to complete development of PCB01 in the first half of 2024 and launch two versions which target large technology companies and general consumers by the end of the year.

Technological Insights, Leadership & Collaboration in the AI Era

Figure 5. Technical Leader Seoyeong Jeong talks about the outlook of the HBM market

During the conference, SK hynix also hosted sessions on key AI memory technologies. These included a presentation by Technical Leader Seoyeong Jeong of HBM Marketing which covered the market outlook for HBM and emphasized the importance of memory solutions in the AI era.

Through its activities at GTC 2024, SK hynix underlined its leadership in the AI memory sector. Looking ahead, the company will continue to realize technological breakthroughs and advance AI technology.

The post NVIDIA GTC 2024: SK hynix Presents the Future of AI Memory Solutions first appeared on SK hynix Newsroom.