Micron’s memory boom: Is the cyclicality really ending?
In Breaking Analysis 322, we shared with you that our forecast of a $1 trillion silicon business arrived two years earlier than we had forecast. The reason for our 24-month miss was the bottleneck had moved from graphics processing units to memory. Micron Inc. gave us a new datapoint in its third-quarter earnings print. It […] The post Micron’s memory boom: Is the cyclicality really ending?…
Micron's latest third-quarter earnings show DRAM prices surged by over 60% sequentially, while shipped bits remained stagnant. This development prompts the question: is this temporary undersupply, or is there a structural shift in memory economics? We explore this with Micron's President and CTO, Scott DeBoer, and delve into the implications for artificial intelligence.
Micron's recent adjustments in risk-sharing agreements, higher customer costs, and the importance of memory in AI performance shape the evolving landscape. We analyze data from Qualitate, examining the supply chain, fabless designers, and IDMs grappling with high-bandwidth memory (HBM) and CoWoS constraints. The divergence between average selling prices and bit growth highlights a unique characteristic of today's semiconductor market, where memory ASPs and bit volume growth no longer move in tandem.
While memory now determines AI system utilization, escalating prices can also prompt customers to defer spending or modify infrastructure. As pricing pressures continue, we question how long this pricing power and profitability will persist as supply tightens and customers adapt to the changing economic dynamics of AI.
Written by urgent.news from SiliconANGLE's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
