Samsung Develops 500-Layer V11 NAND for NVIDIA CMX

Samsung is developing V11 NAND with 500 layers to power NVIDIA's Context Memory Storage, addressing AI GPU data bottlenecks.

NVIDIA is utilizing Context Memory Storage (CMX) to resolve Key-Value cache bottlenecks in AI GPUs, with Samsung supplying critical V-NAND technology. Samsung currently provides 286-layer V9 NAND, which accounts for 60 percent of its 100,000-unit production capacity for NVIDIA. While mass-producing V10 NAND—offering 50 percent higher density than V9—Samsung is researching V11 NAND featuring 500 vertically-stacked layers. Each CMX module for AI servers can house 576 SSDs, reaching a total capacity of 9,600TB. As CMX storage demand is projected to exceed 100 million terabytes by 2026, Samsung expects its NAND shipments to reach 250 million terabytes. This growth is anticipated to drive Samsung's 2026 profits beyond its cumulative earnings from the past 40 years.


Modder Cools RTX 4060 With 2.5kg Aluminum Block

Costco Shopper Finds iBuyPower PC Deal for $1,220

Intel and SK Hynix Deny Ohio Factory Takeover

Google Developing Frozen v2 AI Server Chip

Huawei Kirin 9030 Features 32.5nm Metal Pitch

Z.ai Activates 1-Gigawatt Domestic Chip Data Center

Nvidia Debuts DLSS 5 for Real-Time 4K Performance

Intel Unveils 8000 MT/s DDR5 RDIMMs for Xeon 6

NVIDIA Unveils Vera Data Center CPU at SIGGRAPH

TSMC to Raise Chip Production Prices in 2027

MSI Debuts GeForce RTX 5080 INSPIRE 3X OC BLACK

Yeston and Maxsun Launch RTX 5060 with 16-Pin Power

Modders Unlock Blackwell GDDR7 Temperature Monitoring

Samsung Electronics America Moves HQ to Texas

DIY 3D-Printed Steam Machine Shell Replaces dbrand

AMD Zen 6 Medusa Point 10-Core CPU Leaks on Geekbench

OneXPlayer X2 Mini Pro Handheld Launches

SK Group Chairman Warns of High Memory Prices

Microsoft and AMD Partner for Helios AI on Azure

NVIDIA Previews RTX Spark Windows on Arm Drivers