Skip to content

Samsung Semiconductor Joins MemCon 2023 as Founding Member; Showcases Latest Memory Solutions for AI

  • mail
Today at MemCon 2023, industry leaders in memory and technology gathered to discuss insights and vision into the latest advancements, opportunities, and challenges of innovation in the AI era. This is MemCon’s inaugural event in Mountain View, CA, and Samsung Semiconductor, a leader in memory, is a proud Founding Partner.
Vice President Han Jin-man of the Americas Division (DSA) is presenting on stage.
Vice President Han Jin-man of the Americas Division (DSA) is presenting on stage.
Samsung Electronics’ EVP and Device Solutions America President Jinman Han kicked the event off and stressed the need for breakthrough memory innovation or running the risk of remaining ‘old world.’ Potential innovations include memory-centric architecture and memory-as-a-service. Samsung predicts new business models to emerge as it continues to innovate and improve memory technology. Additionally, Jin-Hyeok Choi, EVP of Device Solution Research America – Memory, Samsung Electronics, presented a keynote speech on “Memory Innovation in the Era of Data-centric Computing.” He shared Samsung Semiconductor’s approach to overcoming a wide range of DRAM and NAND Flash Memory Wall1 problems and Samsung’s vision to lead memory innovation in the semiconductor industry.
Vice President Choi Jin-hyuk of the Americas Memory Research Institute (DSRA-Memory) is giving a presentation on stage.
Vice President Choi Jin-hyuk of the Americas Memory Research Institute (DSRA-Memory) is giving a presentation on stage.
The audience is listening to the presentations.
The audience is listening to the presentations.
Technology Showcase Samsung Semiconductor also showcased memory products such as HBM-PIM and CXL-based PNM that support hyper-scale AI models, SmartSSD, Memory-Semantic SSD, and Memory Expanders.
An illustration explains the structure of HBM-PIM.
An illustration explains the structure of HBM-PIM.
The product image of the CXL-based PNM solution is shown.
The product image of the CXL-based PNM solution is shown.
HBM-PIM and CXL-PNM are known for their powerful functionalities that dramatically improve the performance of GPU accelerators when implemented in AI applications, compared to existing solutions.
Samsung Semiconductor products are on display.
Samsung Semiconductor products are on display.
The 2nd Gen SmartSSD can reduce processing time for data operation and energy consumption. Samsung’s SmartSSDs can provide 20x faster random read and response speed than the general SSDs, making them an excellent fit for AI and Machine Learning. • Samsung’s Memory Expander (MXP) and Memory-Semantic SSDs also utilize the next-generation interface, CXL. Memory Expanders can expand memory bandwidth and capacity to process vast amounts of data much faster. Memory-Semantic SSDs improve the random read and response speeds up to 20 times more quickly.
Various stakeholders from different industries and fields are engaged in discussions.
Various stakeholders from different industries and fields are engaged in discussions.
Visitors are admiring Samsung Semiconductor's products.
Visitors are admiring Samsung Semiconductor's products.
Next-generation memory solutions usually go through the development process while the standardization process is also in progress. That is why continuous collaboration with customers is the key to product commercialization and expansion of the industry ecosystem. MemCon 2023 served as an excellent arena for the industry to gather in one place and have in-depth discussions on the latest memory tech, providing attendees an ideal opportunity to understand the need for next-generation memory products. Samsung Semiconductor continues to advance memory technology for the AI era with relentless innovation and looks forward to collaborating to bring new solutions to the market. 1 Memory Wall describes a performance bottleneck caused by an increasing gap between processor speed and the memory access time. This issue limits system efficiency as the processor spends more time waiting for data from memory.

Explore more stories

For more latest news in the Semiconductor Newsroom, click the button below.

Go to Samsung Semiconductor Newsroom