Micron CEO Blames Price Wars for Memory Chip Shortage - The Tech Buzz

Micron Technology CEO Sanjay Mehrotra recently stated that the memory chip industry's inability to meet the surge in AI demand stems from years of customer pricing pressure that deterred suppliers fro... Read original →

Micron’s candid remarks signal a structural inflection in the memory industry. Years of hyperscaler-driven price suppression starved DRAM players of capex, leaving HBM3E/HBM4 capacity critically inadequate for AI’s exponential bandwidth demands. Technically, HBM is now inseparable from advanced packaging (e.g., CoWoS), TSV etching tools, and GPU architectures—making memory the bottleneck across the AI stack. Compliance-wise, U.S. CHIPS Act restrictions bar Micron from expanding advanced nodes in China for a decade, forcing costly HBM investments in Japan and the U.S., elongating lead times. Samsung and SK Hynix will likely accelerate HBM4 development and demand multi-year customer commitments to secure ROI. Over the next 18 months, HBM pricing power will intensify, locking out smaller rivals and forcing AI data centers to redesign their total cost of ownership models—with memory now the primary constraint on compute scalability.