Samsung just became NVIDIA's storage sugar daddy, and the rest of the SSD market is about to get absolutely griefed. The Korean giant is already shipping V9 NAND chips for NVIDIA's Context Memory Storage modules, which offload massive KV cache payloads from AI GPUs, and V10 is ramping into mass production with V11 research targeting a stacked 500-layer design.
Each CMX unit packs 576 SSDs for a jaw-dropping 9,600TB of capacity, and Samsung has dedicated its 100,000-unit V-NAND output with roughly 60 percent flowing to the V9 line alone. That vacuum is already starving consumer and enterprise supply chains, echoing the DRAM shortage the AI boom triggered earlier, except this time the deficit hits flash storage directly.
Industry projections peg CMX storage demand jumping from 35 million TB past the 100 million mark, with Samsung eyeing a 250-million-terabyte NAND shipment target. The silver lining here is that V11's 500-layer architecture could eventually trickle down into cheaper, higher-capacity consumer SSDs, but only if Samsung decides everyone else deserves a slice of the pie.
Each CMX unit packs 576 SSDs for a jaw-dropping 9,600TB of capacity, and Samsung has dedicated its 100,000-unit V-NAND output with roughly 60 percent flowing to the V9 line alone. That vacuum is already starving consumer and enterprise supply chains, echoing the DRAM shortage the AI boom triggered earlier, except this time the deficit hits flash storage directly.
Industry projections peg CMX storage demand jumping from 35 million TB past the 100 million mark, with Samsung eyeing a 250-million-terabyte NAND shipment target. The silver lining here is that V11's 500-layer architecture could eventually trickle down into cheaper, higher-capacity consumer SSDs, but only if Samsung decides everyone else deserves a slice of the pie.
