A Better Newspaper

## Solidigm – AI Storage and the Memory Supercycle **Solidigm** (formerly Intel's NAND business, acquired by SK Hynix) is positioning itself as a key player in addressing the **storage and memory bottleneck** emerging as a critical constraint on AI inferencing performance. ### The AI Storage Problem As AI inference workloads scale, GPUs increasingly sit idle waiting for data—a phenomenon analysts call the "memory wall" or "memory bottleneck." Solidigm is targeting this gap with: - High-density NVMe SSDs optimized for AI data pipelines - Ecosystem partnerships (notably with VAST Data) to integrate storage with AI software stacks - Technologies addressing the gap between GPU memory bandwidth and storage throughput ### The Memory Supercycle Industry analysts have identified a "memory supercycle"—a period of sustained demand growth for both DRAM and NAND flash driven by AI workloads. Solidigm's positioning within this cycle has implications for: - **Enterprise procurement**: Storage is no longer a commodity purchase; AI-optimized storage commands premium pricing - **Supply chain risk**: SK Hynix/Solidigm concentration in NAND supply creates geopolitical risk given US-China tensions and South Korea's semiconductor position - **Competitive dynamics**: Competes with Samsung, Micron, and Western Digital in AI-optimized storage ### VAST Data Partnership Solidigm's partnership with **VAST Data** (a high-performance storage software company) creates a vertically integrated solution targeting AI factory deployments, combining Solidigm's hardware with VAST's data platform. ### Watch For - Solidigm product announcements at Flash Memory Summit and other storage events - SK Hynix earnings calls for memory supercycle demand signals - Enterprise deals combining Solidigm hardware with AI software platforms - US export control developments affecting SK Hynix/Solidigm supply chains