AI has transformed the memory economy: Micron's CEO announces a structural shift in the market
The semiconductor memory industry stands on the brink of a fundamental transformation. Micron CEO Sanjay Mehrotra has stated that the era of cyclical booms and busts, which defined the economics of this market for decades, is becoming a thing of the past. In the new reality, memory is becoming a strategic resource essential for the functioning of artificial intelligence systems.
Speaking at the site of a new semiconductor plant in Boise, Idaho, Mehrotra emphasized that artificial intelligence has radically changed the nature of demand. Customers no longer view memory as a commodity, choosing a supplier solely by price. Now, chips are tailored to specific processors and architectures, and the performance of an entire server directly depends on the quality and characteristics of the memory.
"I call memory the strategic infrastructure of the AI era," he stated. "Today, it is impossible to imagine AI without memory. AI-based systems require more and more memory."
Numbers confirm the shift
Micron's financial results vividly illustrate this turning point. Revenue for the third fiscal quarter soared to $41.46 billion, compared to $9.30 billion in the same period last year. Gross margin jumped to 84.6% from 37.7% a year earlier, and in the current quarter, the company expects to reach around 86%.
The main bottleneck now is production capacity. Mehrotra admitted that factories cannot keep up with demand: data centers alone require approximately 50% more chips than Micron can produce. The company is already investing $250 billion in the construction of two new plants and research within the United States.
Further growth in demand, according to Micron's head, will be driven by manufacturers of autonomous vehicles, robots, and household appliances with built-in AI.
Risks and challenges
However, not everyone shares this optimism. In June, a class-action lawsuit was filed against Micron, Samsung, and SK Hynix: the plaintiffs accuse the manufacturers of collusion and artificially inflating DRAM memory prices. Additionally, Chinese companies such as CXMT are ramping up production, which could bring back fierce price competition—the very thing the industry, according to Mehrotra, has moved away from thanks to AI.
My analysis: Mehrotra's statements are not just corporate optimism, but a marker of a deep shift. The structure of demand is indeed changing: AI workloads require not just volume, but customized solutions with high bandwidth, which reduces price elasticity. However, history teaches us: high margins always attract new players and provoke overproduction. The question is not whether cycles will return, but how deep the next correction will be when the market becomes saturated.