- DapuStor Unveils 245TB PCIe Gen5 QLC SSD for AI Workloads
- The firm joins other vendors targeting hyperscalers with ultra-capacity flash memory.
- The move reflects how artificial intelligence is driving demand for dense NAND storage.
AI workloads are driving demand for much larger SSDs in hyperscale data centers, and in response, DapuStor has developed announced PCIe Gen5 QLC SSD which scales up to 245 TB.
The 122TB versions of the drive are already in customer use as the high-capacity SSD is designed for AI data lakes, vector databases, and large storage pools where data remains online and is accessed frequently.
These workloads generate large volumes of attachments, logs, and videos during AI training and inference. This data is re-accessed rather than archived, which is why dense flash memory is a better choice than mechanical drives.
DapuStor joins the select group
DapuStor SSD uses QLC NAND. By maintaining four bits per cell, it increases the capacity of a single wafer compared to TLC, improving density and cost efficiency.
Previously, concerns about QLC durability and performance were addressed through controller design, firmware management, and data placement methods.
This announcement makes DapuStor the eighth vendor to introduce an SSD with a capacity of approximately 245 TB.
Kioxia showed LC9 SSD with a capacity of 246 TBBye Sandisk introduced a model with a capacity of 256 TB for AI workloads. Solidigm has confirmed plans for 245 TB drivesAnd Micron unveils 122 TB PCIe Gen5 SSDs as part of efforts to reduce dependence on hard drives.
Huawei approached the problem differentlyCombining high-capacity SSDs with controllers to reduce dependence on expensive HBMs in AI systems.
Samsung has outlined roadmaps is beyond its current capabilities, and its main competitor, SK Hynix, teased PS1101native 245TB PCIe Gen5 enterprise drive.
The move to higher-capacity SSDs reflects changes in the way artificial intelligence stores data. Training sets, output results, and logs are accessed frequently, favoring flash memory that handles random I/O while using less rack space and power.
These SSDs are clearly not for consumers or even typical businesses. Platform requirements, price, and scale limit their use in hyperscale environments where flash memory becomes a critical resource.
Follow TechRadar on Google News. And add us as your preferred source to get our expert news, reviews and opinions in your feeds. Be sure to click the “Subscribe” button!
And of course you can also Follow TechRadar on TikTok for news, reviews, unboxing videos and get regular updates from us on whatsapp too much.

:quality(85):upscale()/2024/12/18/996/n/1922794/18881067676352e1846c34.55470252_.png?w=150&resize=150,150&ssl=1)




