In 2025, the diverse storage demands from AI applications are comprehensively reshaping the entire industry chain. From large - scale data centers to cutting - edge edge computing, from popular smartphones to advanced autonomous driving, the demand for memory exhibits four core trends: large capacity, high bandwidth, low latency, and low power consumption. International giants and domestic manufacturers are accelerating their deployments and launching innovative products tailored to different AI workloads. elexcon2025, the Shenzhen International Electronic Show, has compiled some manufacturers and their representative products worthy of attention:

International Manufacturers: Technological Iterations Leading High - Performance Storage

1. Samsung Electronics

•HBM4 Customized Solution: With a single - chip capacity of 48GB, it supports customer IP integration. The bandwidth density is 30% higher than that of HBM3, making it suitable for heterogeneous computing scenarios in GPUs and AI training chips. 

•DDR5 RDIMM 32GB: Each chip has a capacity of up to 32GB, supporting a maximum of 256GB DDR5 memory configuration. With a speed exceeding 12.8GB/s, it meets the computing power requirements of AI servers. 

•UFS4.1 Storage Solution: It enables smartphones to run AI models with millions of parameters. The dedicated 8TB SSD balances high - capacity and heat - dissipation design. Its large - capacity SSD product line already covers 16TB to 256TB, providing underlying support for AI training and data centers.

2. Micron Technology

•HBM3E Memory Module: Stacking 16 - layer DRAM chips, its bandwidth exceeds 1.5TB/s. Paired with the MR - MUF process for optimized heat dissipation, it is compatible with NVIDIA's Blackwell - architecture GPUs, enhancing AI rendering efficiency. 

•6500 ION SSD: With a capacity of 30.72TB and support for NVMe 2.0, it is suitable for large - language - model (LLM) training and inference, with a 25% increase in energy efficiency ratio. 

•PCIe 5.0 SSD 4600: It has a sequential read speed of 14.5GB/s and a write speed of 12GB/s, with a power consumption of only 7W, suitable for AI inference on PC - side.

3. KIOXIA

•The Eighth - generation BiCS FLASH QLC: With a capacity of up to 2TB, a transmission rate of 3200MT/s, and a 70% increase in write energy efficiency, it is used in enterprise - level SSDs and data centers.

•Automotive - grade UFS 4.0: With a capacity of 512GB and a working temperature range of - 40 ~ 105℃, it meets the data - storage requirements of autonomous driving.

4. SanDisk

•iNAND MC EU711 UFS4.1: With a sequential read speed of 4000MB/s and a write speed of 3000MB/s, and support for a large capacity of 1TB, it is suitable for AI inference on the smartphone side. 

•Sandisk DC SN861 Enterprise - level SSD: With a PCIe 5.0 interface and a capacity of 16TB, it has passed the NVIDIA GB200 certification, accelerating LLM deployment. 

•Automotive - grade SSD AT EN610: With M.2 1620 BGA packaging and support for the SLC mode, it is suitable for the central computing architecture of intelligent driving.

5. Solidigm (a subsidiary of SK Hynix)

•D5 - P5336 Enterprise - level SSD: With a capacity of 122TB, using QLC technology and supporting the PCIe 5.0 interface, its 4K random read performance reaches 2500K IOPS, specifically designed for AI training clusters. 

•D7 - PS1010/D7 - PS1030: With 176 - layer TLC NAND, a sequential read speed of 14.5GB/s, and a maximum capacity of 15.36TB, they are specifically designed for AI servers.

Domestic Manufacturers: Technological Breakthroughs to Capture the Local Market

6. Yangtze Memory Technologies Co., Ltd.

•X3 - 6070 QLC Flash Memory: With a P/E cycle life of 4,000 times, a speed of 2400MT/s, a 70% increase in density, it has significant cost advantages. 

•PCIe 5.0 SSD: Based on the Xtacking 4.0 architecture, it can reach a full - speed read speed of 14GB/s, supporting AI training and inference. 

•Xtacking 4.0 QLC NAND: Each chip has a capacity of 2Tb, with a bit density 2.3 times higher than the previous generation. It supports the mass production of 1TB TLC products and is suitable for large - capacity storage in AI servers.

7. Longsys

•UFS 4.1: With the self - developed controller WM7400, it can support both TLC and QLC NAND Flash simultaneously. The sequential read speed is as high as 4350MB/s, and the random read and write speeds are as high as 750K / 630K IOPS, with a maximum capacity of up to 2TB. It helps edge - side AI products (such as AI smartphones, AI tablets, intelligent vehicles, humanoid robots, etc.) make real - time decisions in complex applications and enjoy a smooth experience. •Self - developed Automotive - grade Storage Controller: Automotive - grade storage products cover UFS, eMMC, LPDDR4x, SPI NAND Flash, etc., establishing a "dual - drive" automotive - grade system of Flash and DRAM.

Since 2024, Longsys has developed controllers for eMMC and SD. In 2025, it released controllers for UFS and USB, with core IP designed and developed independently.

8. GigaDevice

•The Leader in NOR + MCU Driven by AI and Domestic Substitution 

•NOR Flash: In 2021, it had a 23% global market share and was the world's third - largest NOR Flash manufacturer, supporting low - power storage for AIoT devices. 

•LPDDR5 Memory: It is compatible with AI models on the smartphone side, improving local computing power and storage efficiency.

9. Yangtze Wanrun Semiconductor (a subsidiary of Wankun Technology Co., Ltd.)

•High - performance Storage Products: Yangtze Wanrun Semiconductor's SSD, PSSD, eMMC and other products can be widely used in smartphones, computers, servers, data centers and other fields, providing strong support for high - performance computing and massive data storage application scenarios. 

•Green Storage Technology: Through technological innovation, a series of low - power storage products have been launched, further reducing the energy consumption of terminal devices and contributing to sustainable development. 

•Scenario - based Solutions: Aiming at the differentiated needs of different industries, Yangtze Wanrun Semiconductor has carefully created customized storage solutions to help terminal customers quickly implement their product solutions.

10. TWSC®

• •HBM4 Customized Solution: With a single - chip capacity of 48GB, it supports customer IP integration. The bandwidth density is 30% higher than that of HBM3, making it suitable for heterogeneous computing scenarios in GPUs and AI training chips. 

•DDR5 RDIMM 32GB: Each chip has a capacity of up to 32GB, supporting a maximum of 256GB DDR5 memory configuration. With a speed exceeding 12.8GB/s, it meets the computing power requirements of AI servers. 

•UFS4.1 Storage Solution: It enables smartphones to run AI models with millions of parameters. The dedicated 8TB SSD balances high - capacity and heat - dissipation design. Its large - capacity SSD product line already covers 16TB to 256TB, providing underlying support for AI training and data centers.

Technological Trends: How Will Storage Reshape the Future of AI?

•Integration of HBM and CXL Technologies: Through the integrated memory - computing architecture, the combination of HBM4 and the CXL protocol can reduce data transfer latency and improve AI training efficiency.

•Popularization of QLC NAND: QLC SSDs with higher density and lower cost have become standard in AI data centers. Manufacturers such as KIOXIA and Solidigm are driving capacity breakthroughs beyond 100TB. 

•Optimization of Edge - side Storage: Technologies such as LPCAMM2 and UFS4.0 help smartphones and autonomous - driving terminals achieve local AI inference.

The Cluster of Storage Exhibitors at elexcon

It is worth noting that elexcon, the Shenzhen International Electronic Show, has a rich cluster of semiconductor memory exhibitors, including Longsys, Biwin, Skytree Innovation, KINGSHARE, Kingbox, Speed Digital, Vastgoal Information, Dongxin Semiconductor, Ranpo Technology, HIKVISION Storage, Netac, Orient Fusion, ADATA, Qxunlink, Hongxingyu Technology, Comin Semiconductor, Yuxin Semiconductor, Comchipway, Apacer, Chifu Technology, Hengshuo Microelectronics, Ziguang Guoxin, Yangtze Wanrun, Xinghuo Semiconductor, Hengshuo Microelectronics, Kingston, Tigo, TWSC®, Huaxingxin Semiconductor, etc.

Looking back at elexcon2024, the Shenzhen International Electronic Show, exhibitors such as Ziguang Guoxin, Longsys, Dongxin Semiconductor, TWSC®, Yangtze Wanrun Semiconductor, Skytree Innovation, Comin Semiconductor, Qxunlink Technology, Comchipway, HIKVISION Storage, Yuxin Semiconductor, etc. all demonstrated their capabilities, presenting the latest technological advancements in memory and products like SSDs. These cover fields such as consumer electronics, industrial - grade applications, automotive - grade applications, and high - performance AI scenarios.

Ziguang Guoxin

At the elexcon2024 Shenzhen International Electronic Show, Ziguang Guoxin showcased a number of cutting-edge products and innovative solutions, targeting several key fields such as automotive electronics, artificial intelligence, high-performance computing, industrial control, the Internet of Things, and consumer electronics.

Ziguang Guoxin's new-generation automotive-grade low-power dynamic random access memories, such as LPDDR4/LPDDR4X, and many other high-reliability automotive-grade chip products and solutions with independent intellectual property rights can be applied to the fields of automotive electronic control systems, including lidar, vehicle networking terminals, head-up displays, liquid crystal instrument panels, vehicle dynamic control systems, dash cams, and driver monitoring systems.

Aiming at applications in fields such as large-scale artificial intelligence models and high-performance computing (HPC), Ziguang Guoxin's solutions of "achieving large bandwidth and large-capacity embedded storage at the chip level using embedded DRAM (SeDRAM®) technology" and "expanding memory capacity and bandwidth at the system level using CXL technology" have achieved a memory access bandwidth of dozens of terabytes and a storage capacity of dozens of gigabytes.

Comin Semiconductor

At the site of the elexcon2024 exhibition, with the theme of "Moving Forward with the Core, Intelligent Storage without Borders", Comin Semiconductor launched three new self-developed storage products: the Star River Core Sprite - a self-developed controller eMMC embedded storage chip, the Swift Shadow Core Little Jupiter - a self-developed controller microSD mobile memory card, and the Portable Storage Core Little King Kong PSSD - a portable magnetic suction mobile solid-state drive.

In addition, for AI glasses applications, Comin Semiconductor has introduced the ePOP embedded storage chip. This chip integrates eMMC and LPDDR within one package, which is not only smaller in size but also more powerful in performance. By being vertically mounted on the SoC, the ePOP saves 60% of the space, with the minimum thickness being only 0.8mm. Currently, this product has been widely applied to smart wearable devices such as AI smart glasses and smart watches.

HIKVISION Storage

HIKVISION Storage showcased its consumer-grade and industry-grade storage products. With the embedded product line as the core, the exhibited products are mainly targeted at the fields of servers, in-vehicle applications, industrial control, and consumer electronics.

Currently, HIKVISION Storage's embedded storage products have covered product series such as eMMC, LPDDR, DDR4, SPI NAND, and Parallel NAND, meeting the demands of segmented markets such as power, network communication, transportation, mobile terminals, and smart homes.

Netac Technology

Netac Technology displayed products in multiple categories, including SSDs, DDR memories, mobile solid-state drives, flash drives, and memory cards, covering numerous application scenarios such as daily office work, entertainment gaming, business trips, and travels. Among them, star products such as the Jueying NV7000, Jueying RGB memory, and Z2S mobile solid-state drive have attracted much attention.

Comchipway

Comchipway comprehensively showcased its self-developed embedded storage chips, including eMMC5.1 products. The eMMC memory has achieved mass production in fields such as consumer electronics, industrial control, and in-vehicle applications. Its products have entered the supply chains of brands such as ZTE, Jiulian Technology, Hisense, Konka Group, Star-net Communication, Inspur Group, and Shenzhen Bolong Industrial Co., Ltd. As of July 2024, the cumulative sales volume has exceeded 50 million units.

Yuxin Semiconductor

Yuxin Semiconductor exhibited a variety of storage products, including high-performance embedded storage eMMC based on the latest 3D NAND of Yangtze Memory Technologies Co., Ltd., SPI NAND, memory chips such as DDR3L/DDR4/LPDDR4x, raw NAND flash memory chips, high-speed micro SD flash memory cards, consumer-grade and enterprise-grade SATA/mSATA/M.2 SSDs, PCIe3.0/4.0 SSDs, etc., as well as various storage solutions.

Skytree Innovation

Skytree Innovation showcased high-performance embedded storage chips represented by UFS 3.1, LPDDR5, eMMC5.1, and ePOP, along with application cases.

Among them, the UFS3.1 is designed for AI and flagship smartphones. It has a sequential read speed of up to 2100MB/s and a sequential write speed of up to 1700MB/s, with a theoretical bandwidth of 2.9GB/s. It has dimensions of only 11.5×13.0×1.0mm and features high speed, stable compatibility, durability, and reliability. The LPDDR5 has a transmission rate of 6400Mbps, a bandwidth of up to 51GB/s, and a working voltage as low as 0.5V. Compared with LPDDR4/4X, its overall performance is improved by 50% and power consumption is reduced by up to 30%, combining the advantages of high performance and low power consumption. The eMMC5.1, based on Skytree Innovation's self-developed firmware algorithm, has a sequential read speed of up to 320MB/s and a sequential write speed of up to 200MB/s, with a large storage capacity of 256GB. It is characterized by small size, high integration, and high reliability.

Qxunlink Technology

The high-end storage products displayed by Qxunlink Technology cover consumer-grade, industrial-grade, automotive-grade, and enterprise-grade categories. The enterprise-grade products include U.2 PCIe 5.0 QLC SSD, ES33H01 U.3 NVMe SSD, and DRR4 R-DIMM. In terms of AI products, the Qxunlink AI Training and Inference All-in-One Machine is mainly composed of core components such as the Qxunlink Tianyi AI Expansion Card, enterprise-grade solid-state drives, and server memories, including the Qxunlink Tianyi AI Expansion Card, DDR5 R-DIMM, SATA/NVMe eSSD, etc.

Shenzhen International Electronic Show

elexcon2025 Embedded Exhibition

The elexcon2025 Shenzhen International Electronic Show, hosted by UBM Creative Exhibition, will be held at the Shenzhen Convention and Exhibition Center (Futian) from August 26th to 28th, 2025. The theme of the fair is "AII for AI, AIl for GREEN: Providing Comprehensive Technical and Supply Chain Support for AI and the Dual Carbon Goals". The exhibition will centrally display cutting-edge products, technologies, and solutions in areas such as AI and computing power chips, storage, embedded systems and AIoT, power supplies and energy electronics, and the Chiplet heterogeneous integration ecosystem. It will comprehensively cover popular fields such as artificial intelligence, new energy vehicles, industrial automation, consumer electronics, the Internet of Things, healthcare, and robotics, and promote technological innovation and industrial development. Currently, booth reservations are in full swing. You are welcome to sign up!

•Display Scope: Embedded and Edge-side AI, Edge AI Computing Platforms, AI Industrial Computers, Embedded Processors/SoC/MCU/MPU, High Computing Power Processor Core Boards, High-speed Storage, RISC-V and Open Source, Smart Displays/HMI, Industrial Power Supplies, Wireless Communication, OS Operating Systems, Software and Tools, Other Components, Solutions, etc. 

 •Target Audience: Professionals from industries such as artificial intelligence, machine vision, industry, automotive, communication, rail transit, healthcare, the Internet of Things, as well as those from universities, colleges, and scientific research institutions.

The 7th China Embedded Technology Conference

Forum 1: Integration of AI and Edge Intelligence

Topics: Innovation in embedded AI chip architecture, deployment of lightweight algorithms, real-time decision-making applications at the edge (such as edge-side AI).

Forum 2: Open Source Ecosystem and Opportunities for Domestic Development

Topics: Applications of RISC-V in IoT/automotive-grade scenarios, joint construction of China's open-source chip ecosystem, competition and cooperation between the RISC-V and ARM ecosystems, the Raspberry Pi ecosystem, and the RT-Thread operating system.

Forum 3: Industrial Control and Applications

Topics: Industrial and automotive chips, industrial embedded basic software, motor control and drive, industrial wired and wireless networks, functional safety and IoT security, industrial intelligent terminals, display technology, and innovative applications of low-power MCUs.

Forum 4: AIoT Intelligent Internet of Things System

Topics: AI architecture for end-edge-cloud collaboration, the trend of customized AIoT chips, integrated applications in vertical industries (smart healthcare, smart retail, smart home), industrial IoT, etc.

Forum 5: Embedded Systems and Applications of Embodied Robots

Topics: Real-time motion control algorithms, applications of edge AI in robot decision-making (such as visual servo control), the Robot Operating System (ROS 2.0) and real-time performance optimization, design of modular robot hardware architecture.

Forum 6: Storage Technology Forum

Topics: Reliability of automotive-grade/industrial-grade storage devices, applications of new non-volatile memories, integration of storage and computing, optimization of time series databases (TSDB) at the edge, secure data erasure and privacy protection at the edge side, and the practice of storage-computing chips in AI inference.