Micron Technology's stock has surged over 500 percent in the past year, but the gains now reflect a structural bet on sustained AI memory demand that requires tens of billions in foundry buildout—and offers limited margin for error if architectures shift.
In fiscal Q3 ending May 28, Micron posted $41.46 billion in revenue, up from $23.86 billion the prior quarter and $9.30 billion year-over-year. Gross margin exceeded 84 percent. Management guided to approximately $50 billion in fiscal Q4 revenue with gross margin near 86 percent.
The supply picture appears locked. Micron stated months ago that all high-bandwidth memory production for the current year had sold through long-term agreements. More recently, the company announced HBM products are sold out through 2027, with demand for HBM4 extending beyond that period.
But the capital commitment required to fulfill these multiyear contracts is substantial and inflexible. In fiscal Q3, Micron invested $7.1 billion in capital expenditures. The company maintained over $30 billion in cash and marketable investments at period end, supported by strong operating cash flow. Yet the buildout of AI-focused chip foundry and packaging capacity now underway represents a bet on demand durability over years, not quarters.
The risk is structural. If AI architectures shift materially—toward different memory hierarchies, chiplet configurations, or competing technologies—or if pricing compression accelerates, the new capacity still generates carrying costs and opportunity costs. A memory fab takes years to amortize. A cycle downturn while utilization remains low compounds the problem.
Memory supply constraints do favor near-term pricing power. New production capacity typically requires 18 to 36 months to come online; demand for AI accelerators shows no immediate signs of cooling. These timelines have historically protected memory producers during growth phases.
Yet semiconductor memory is inherently cyclical. The AI supercycle does not eliminate this dynamic—it only delays the reversion. Micron's current valuation already prices in years of elevated demand and pricing. Investor expectations are now so high that in-line results can trigger selloffs.
As Micron approaches its next earnings report, the market will focus on customer agreements, product pricing trends, and capital allocation decisions. Each signal will be parsed for evidence of demand durability. The core question remains: whether the company's CapEx commitment is proportional to demand or a generational bet that the AI cycle behaves differently than memory cycles of the past.
