AI Memory Shortage Ignites Micron’s Record Results in Q3

Micron Technology (NASDAQ:MU) reported an exceptional fiscal third quarter with record revenue, gross margin, and EPS above the high end of guidance. Revenue reached $41.5 billion, up 74% sequentially and 346% year over year, while fiscal Q3 free cash flow reached a quarterly record of $18.3 billion.

The company said data center revenue exceeded a $25 billion annualized run rate in the quarter, and data center SSD revenue exceeded a $5 billion run rate, more than doubling sequentially.

Discussion centered on the growing role of memory in AI infrastructure.

Management emphasized that AI system performance depends on memory subsystem performance and capacity, making memory a key asset for enterprise and product roadmaps. Micron also highlighted 16 customer agreements, or SCAs, designed to provide contracted supply assurance and improve visibility for technology and manufacturing investments.

Why It Matters: The remarks showed memory and storage function as key enablers of enterprise AI infrastructure. The company described AI systems’ dependency on memory performance, which affects planning for data center architecture and edge deployment. Supply agreements and manufacturing expansion also indicate that access to advanced DRAM, NAND, and HBM may remain a planning issue while AI demand continues to pressure supply.

• AI Elevates Memory’s Role: CEO Sanjay Mehrotra said, “AI system performance is architecturally dependent on memory subsystem performance and capacity,” creating more differentiation opportunities for Micron. He said this has “elevated the role of memory in the AI world to a strategic asset,” with customers treating advanced memory access and committed supply as essential to product roadmaps.

• Customer Agreements Support Supply Planning: Micron said it has signed 16 SCAs with customers across major end markets, covering roughly 20% of DRAM volume and one-third of NAND volume over the agreement periods. The company described the SCAs as take-or-pay agreements with binding purchase commitments, giving customers supply assurance while improving Micron’s visibility into demand. Management said the agreements also support technology collaboration and investment planning while AI-driven demand continues to exceed industry supply.

• High-Bandwidth Memory growth supports AI infrastructure: The company said HBM4 12-High volume ramp is tracking twice as fast as HBM3E 12-High, and the company has already shipped more than $1 billion in HBM4 revenue. Management described HBM as vital for AI use cases across data center and edge environments, while noting that HBM consumes major wafer capacity and pressures non-HBM supply. Micron said its HBM strategy is to keep market share close to its DRAM share so it can support AI customers and other end markets that require non-HBM supply.

• Agentic AI Expands Infrastructure Needs: Agentic AI is changing data center infrastructure, increasing demand for compute and storage systems that support control-plane activity, program execution, and context stores. The company also pointed to AI context memory storage and hard-drive replacement opportunities that expand the addressable market for SSDs.

• Physical AI Drives Future Memory Demand: Executives said physical AI applications are creating demand for high-bandwidth memory and storage that supports real-time perception and control. Management said advances in simulation and integrated system design are increasing adoption in autonomy. Mehrotra noted that humanoid robots carry 10x the memory of an average L2+ vehicle, and the company expects a multi-decade memory demand cycle to begin later this decade.

Go Deeper -> Micron Technology’s Earnings Report – MarketBeat

---

Our Latest CIO Field Notes

---