Micron Technology (NASDAQ: MU) CEO Sanjay Mehrotra declared on Thursday that artificial intelligence has fundamentally reshaped the memory chip industry, a sector historically defined by violent boom-and-bust cycles. "Today, there is no AI without memory. AI systems demand more memory, higher-performance memory, and lower-power memory," Mehrotra said during an interview with Jim Cramer on "Mad Money." "The value equation for memory has completely changed."
His remarks came against the backdrop of a massive semiconductor manufacturing complex under construction, located adjacent to Micron's headquarters in Boise, Idaho. The site is part of a planned $250 billion investment in U.S. manufacturing and R&D. According to Mehrotra, the Boise campus alone will eventually house two wafer fabrication plants, each the size of ten football fields, with enough steel reinforcement in a single facility to circle the Earth twice if laid end-to-end. The first Boise fab is expected to begin wafer production in mid-2027.
The scale of this investment reflects how profoundly Mehrotra believes AI is rewriting the outlook for the memory market. The industry has historically been cyclical, with surging demand prompting capacity expansion, which in turn leads to oversupply and falling prices. However, Mehrotra, an engineer by training with over four decades in the chip industry who also co-founded SanDisk (SNDK.US), points out that AI is generating a far more durable source of demand.
The AI era not only requires compute models but also demands the continuous movement, preservation, and low-latency retrieval of massive state data. While HBM and DRAM handle the highest-bandwidth working sets near GPUs, model weights, training datasets, checkpoints, vector databases, RAG knowledge bases, inference logs, agent long-term memory, and layers of KV cache cannot all reside permanently in expensive, capacity-limited HBM. This is where NAND-based storage products like enterprise NVMe SSDs begin to assume the role of a high-performance persistent layer between HBM/DRAM and HDDs or object storage.
He emphasized that this opportunity extends far beyond the data center. Mehrotra expects autonomous vehicles, robotics, and AI-enabled consumer devices to place increasingly higher demands on storage capacity over the coming years. "Memory is now indispensable," Mehrotra stated. "That is why I call it strategic infrastructure for the AI era."
Mehrotra also noted that this demand is changing how customers perceive the value of memory. In the past, customers would solicit bids and choose the cheapest supplier. Now, memory must be co-designed and co-developed with processors and systems. This makes memory a critical component of overall system performance rather than a replaceable commodity. "We engage with customers very early in their development cycle, even earlier than before," he said. "Our customers recognize the value of memory because it is essential to designing products that drive their business growth."
He stated that Micron currently cannot produce enough chips to satisfy demand. "Across all end markets, our customers will buy everything we can manufacture," Mehrotra added, noting that data center customer demand is currently running about 50% above what Micron can commit to supply. To gain better visibility into demand trends, Micron is enhancing predictability through long-term customer agreements, a significant shift for an industry historically exposed to spot market price volatility. At the company's most recent earnings call in late June, Micron announced it had signed five-year strategic agreements with 16 customers. Mehrotra said more agreements have been signed since then. "They have committed to purchase volumes," Mehrotra said. "This provides us with a stable demand foundation."
Comments