Memory Market Transformation: Servers and Automotive Emerge as Key Drivers

At the recent 2026 IC China industry conference, Luo Xiaodong, Vice President of ChangXin Memory Technologies (CXMT), provided a detailed analysis of the evolving global memory demand landscape. He emphasized that the industry is moving beyond a one-size-fits-all approach into a highly differentiated 'Stratified Era'.

Server Market: Dominates with Over 50% Share, Trends Toward High-End

Luo revealed that memory demand from the server market now accounts for more than 50% of total demand, making it the dominant force. Growth in this sector is not just quantitative but qualitative.

  • Surge in High-Performance Chip Demand: Complex workloads like AI, big data analytics, and cloud computing are driving urgent need for high-bandwidth, low-latency memory.
  • High-End Storage Solutions in Demand: Servers are shifting from general-purpose to accelerated computing, requiring more specialized and efficient memory solutions.

This necessitates memory suppliers to pivot their R&D focus towards meeting these high-end, specific server requirements.

Automotive Electronics: The Next Growth Frontier, 30% CAGR Projected

Beyond servers, smart vehicles are seen as the next golden track for the memory industry. Luo projected that driven by advanced infotainment, autonomous driving, and in-vehicle systems, demand for automotive-grade storage and chips will maintain a high compound annual growth rate of around 30% until 2030.

"Storage capacity per vehicle is increasing rapidly," he added. "Future cars will not merely be transportation tools but mobile intelligent terminals with powerful computing capabilities, posing new challenges for memory reliability, endurance, and performance."

Beyond Conventional Thinking: The Five Core Dimensions of Memory

Confronted with diversified market needs, Luo urged the industry to move past the traditional framework focusing solely on capacity and speed. He proposed that modern memory performance and value should be evaluated through five core dimensions:

  1. Capacity: The fundamental storage space.
  2. Bandwidth: Data throughput capability.
  3. Write Rate: Efficiency of persistent data writing.
  4. Access Granularity: The smallest unit of data read/write, affecting fine-grained operation capability.
  5. Latency: Response time for data access.

This multi-dimensional model provides a more precise roadmap for the design, selection, and optimization of future memory products, signaling an evolution from generic components to scenario-specific customized solutions.