Stay Informed. Stay Ahead

Orld – Categories

Kioxia Unveils 5TB Flash Memory Module for AI Applications

Kioxia, a global leader in memory solutions, recently unveiled a groundbreaking prototype large-capacity flash memory module crucial for handling extensive artificial intelligence (AI) models. This development is a part of the “Post-5G Information and Communication Systems Infrastructure Enhancement R&D Project” initiated by Japan’s national research agency, NEDO. The flash memory module boasts an impressive five terabytes capacity and a high bandwidth of 64 gigabytes per second.

NOYITO W25Q64 64Mbit 8MByte Flash Memory Module DataFlash SPI Interface Blue

NOYITO W25Q64 64Mbit 8MByte Flash Memory Module DataFlash SPI Interface Blue | $3.99

To overcome the conventional trade-off between memory capacity and bandwidth, Kioxia devised a novel module configuration utilizing daisy-chained connections with flash memories. By integrating high-speed transceiver technology capable of supporting bandwidths of 128 gigabits per second, along with performance-enhancing techniques, Kioxia has successfully enhanced both memory controllers and modules.

NOYITO W25Q16 16M-bit 2M-Byte Memory Module W25Q16BVSIG Serial SPI Flash Module

NOYITO W25Q16 16M-bit 2M-Byte Memory Module W25Q16BVSIG Serial SPI Flash Module | $4.99

The practical application of this advanced memory module is poised to revolutionize digital transformation by facilitating the integration of Internet of Things (IoT), Big Data analytics, and advanced AI processing into post-5G/6G Mobile Edge Computing (MEC) servers and other applications.

NOYITO W25Q32 W25Q64 W25Q128 Flash Memory Module Data Flash SPI Interface (Pack of 5) (W25Q128)

NOYITO W25Q32 W25Q64 W25Q128 Flash Memory Module Data Flash SPI Interface (Pack of 5) (W25Q128) | $13.99

In the era of post-5G/6G networks, the demand for faster speeds, reduced latency, and increased device connectivity necessitates the adoption of MEC servers. These servers process data closer to users, thereby enhancing real-time application performance. With the rising need for sophisticated AI applications, such as generative AI, there is a growing requirement for memory modules with larger capacity and higher bandwidth.

Alinan 10pcs W25Q32 Power Module Large Capacity Flash Storage Module SPI Interface Memory Card BV FV STM32 Code for Repair...

Alinan 10pcs W25Q32 Power Module Large Capacity Flash Storage Module SPI Interface Memory Card BV FV STM32 Code for Repair… | $11.99

Against this backdrop, Kioxia focused on leveraging flash memory to enhance memory module capacity and bandwidth. By successfully developing a prototype memory module with a capacity of 5 TB and a bandwidth of 64 GB/s, Kioxia has demonstrated the feasibility and effectiveness of their approach.

SAMSUNG MUF-256AB/AM FIT Plus 256GB - 300MB/s USB 3.1 Flash Drive, Gunmetal Gray

SAMSUNG MUF-256AB/AM FIT Plus 256GB – 300MB/s USB 3.1 Flash Drive, Gunmetal Gray | $22.99

Kioxia’s achievements in this project include the adoption of daisy-chain connections with controllers and memory boards, enabling high-capacity, high-bandwidth memory modules. The utilization of 128 Gbps PAM4 transceivers and technologies enhancing flash memory performance have further augmented the capabilities of these modules.

32GB USB 2.0 Flash Drive, BorlterClamp Memory Stick Retro Metal Love Heart Key Shaped Thumb Drive

32GB USB 2.0 Flash Drive, BorlterClamp Memory Stick Retro Metal Love Heart Key Shaped Thumb Drive | $10.99

Looking ahead, Kioxia plans to commercialize the outcomes of this research swiftly to align with emerging market trends such as generative AI. The company aims to capitalize on these developments to drive innovation in IoT, big data analysis, and advanced AI processing at the edge.

Ray Sharma, an Industry Analyst with extensive experience in mobile broadband technologies, underscores the significance of Kioxia’s advancements in memory solutions. He emphasizes the impact these developments can have on advancing network management and enabling cutting-edge technologies like generative AI.

📰 Related Articles


📚Book Titles