The semiconductor memory market is undergoing a fundamental shift that challenges the industry's decades-long cyclicality. Micron CEO Sanjay Mehrotra, speaking at the construction site of a new plant in Boise, Idaho, stated that artificial intelligence has turned memory into a strategic resource, permanently changing the economics of this business.
Traditionally, the memory industry developed along a predictable sine wave: rising demand spurred construction of new capacity, overproduction crashed prices, and the cycle repeated. However, as Mehrotra emphasizes, the AI era is breaking this paradigm. Customers can no longer view memory as a commodity, choosing suppliers solely by price. Chips are increasingly designed for specific processors and architectures, and the performance of an entire server depends on the tight coupling of components.
"That's why I call memory the strategic infrastructure of the AI era," said the Micron chief. "Today, it's impossible to imagine AI without memory. AI-based systems require more and more memory."
Numbers confirm the ambitions
Micron's financial results eloquently illustrate this transition. Revenue for the third fiscal quarter soared to $41.46 billion compared to $9.30 billion in the same period last year — a more than fourfold increase. Gross margin reached an impressive 84.6% versus 37.7% a year earlier, and the company forecasts the figure at around 86% for the current quarter.
The key problem now is not demand, but supply. Mehrotra acknowledges that existing plants cannot satisfy customer appetites. Data centers alone require approximately 50% more chips than Micron is capable of producing. The company is already implementing a massive $250 billion investment plan in manufacturing and research across the United States, including the construction of two new facilities.
Looking ahead, according to the Micron chief, demand will only intensify. Among the growth drivers, he cites manufacturers of autonomous vehicles, robotics, and household appliances with built-in AI.
Not so clear-cut
However, not everyone shares this optimism. In June, a class-action lawsuit was filed against Micron, Samsung, and SK Hynix: plaintiffs accuse the manufacturers of collusion and artificially inflating DRAM memory prices. Additionally, Chinese companies such as CXMT are aggressively ramping up output, which could bring back fierce price competition — the very thing the industry, according to Mehrotra, has moved away from thanks to AI.
My take: Mehrotra's statements are not just corporate optimism, but a marker of a shift in technological paradigm. AI is indeed creating structural, not cyclical, demand for memory. However, history knows examples where a "new era" ended in overproduction. Investors should closely monitor not only financial metrics but also the capacity balance — it is this that will determine who emerges victorious from this race.