NOR Flash Next in AI-Driven Memory Crunch

2026-02-09 · Source: Big Data & AI News - EE Times · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure, Emerging Technologies & Innovation · Depth: Intermediate, short

Summary

The NOR flash memory market is experiencing a significant supply crunch, primarily driven by surging demand from AI servers and data centers. The number of NOR devices per server rack has increased dramatically, from 3-5 units to over 30, with NOR content in Nvidia's GB200 NVL72 system potentially exceeding $900 per rack within two years. NOR flash is critical for AI server safe boot, initialization, and firmware storage due to its fast random-access read speeds, high reliability, and low latency. It also supports independent power management for HBM devices and is vital for edge AI in storing operating systems and neural network weights. This increased demand is intensifying competition for production capacity, with reports indicating potential price hikes from major suppliers like Macronix, which is also reportedly scaling down NOR capacity to boost MLC NAND flash production. The industry is also looking towards 3D NOR flash, which promises higher densities and lower latency, but it is still years away from widespread commercial availability.

Key takeaway

For CTOs and VPs of Engineering managing AI infrastructure, the escalating NOR flash supply crunch and potential price hikes necessitate proactive supply chain management. You should evaluate your current and projected NOR flash requirements, explore alternative suppliers, and consider the long-term implications of 3D NOR flash development for future system designs. Prepare for potential cost increases and lead time extensions for critical boot and firmware storage components.

Key insights

AI server demand is causing a NOR flash supply crunch, driving innovation towards 3D NOR technology.

Principles

• High reliability is crucial for AI server boot processes.
• Deterministic read latencies benefit edge AI model storage.

In practice

• Use NOR flash for AI server firmware and boot code.
• Integrate NOR flash for HBM power management.
• Store neural network weights on NOR for edge inference.

Topics

• NOR Flash Memory
• AI Servers
• Memory Supply Chain
• 3D NOR Flash
• Edge AI

Best for: Investor, CTO, VP of Engineering/Data, AI Engineer, AI Architect, AI Operations Specialist

Related on AIssential

• Dell, HPE server supply upended by memory crunch — A historic structural disruption in the memory market is fundamentally reordering global memory allocation towards AI infrastructure.
• Massive AI Storage Demand Creates a New Memory Wall — AI's massive memory demands necessitate new architectures like high-bandwidth flash, optimizing capacity and sequential bandwidth over traditional DRAM.
• AI chips are pushing everything else off TSMC's most advanced production lines — Simultaneous AI accelerator migration to TSMC's N3 process is creating a severe, multi-year capacity crunch.
• Why we’re at the beginning of the AI hardware boom | Caitlin Kalinowski (ex–OpenAI, Meta, Apple) — The next frontier for AI is the physical world, driving a critical re-industrialization effort and new hardware development paradigms.
• Why the Memory Crunch Is Almost Impossible to Solve — The global memory-chip shortage is intractable due to concentrated production, long factory lead times, and surging AI demand, compounded by geopolitical constraints.
• 🔮 Exponential View #568: The labs are rationing. Did you notice? — The AI industry faces a severe compute crunch, impacting growth and resource allocation for major tech companies.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Big Data & AI News - EE Times.