Micron Technology reported record revenue, gross margin, and earnings for its fiscal third quarter on Wednesday, sending shares up about 18% in premarket trading Thursday and resetting the AI memory trade that had wobbled earlier in the week.
The bounce followed a sharp sell-off. On Tuesday, the Philadelphia Semiconductor Index had its second-worst day of the past year, and Micron suffered its worst session since the April 2025 "Liberation Day" rout. The company's results answered the market's immediate question: AI-driven memory demand isn't cracking.
AI memory scare meets its match
Micron and South Korea's SK Hynix had been two of the cleanest plays on the AI memory boom this year, both handily outperforming the broader chip index before this week's dip. Tuesday's decline looked like a stress test, but the earnings report put the trade back on firm footing. The company's profit machine showed no signs of slowdown.
The numbers behind the rally
Revenue reached $41.5 billion, well above expectations. Adjusted earnings came in at $25.11 per share. Gross margin hit 84.9%, more than double the year-ago figure and the highest in data going back to 1990. The company expects that margin to rise again this quarter, to roughly 86%.
That gross margin figure is the clearest signal. It shows how much of each sales dollar Micron keeps after production costs, and it's now keeping more than ever. "AI system performance depends on memory performance and capacity," Micron said, framing memory as a strategic asset rather than a commodity add-on.
Customer agreements lock in demand
The longer-term story is in the 16 strategic customer agreements Micron disclosed. These include take-or-pay commitments that run for several years - customers agree to buy set volumes or pay anyway. Fourteen of the deals represent about $100 billion of minimum contracted revenue over their remaining term, backed by $22 billion in cash deposits and related commitments.
For a business historically prone to boom-and-bust cycles, that changes the risk profile. AI customers aren't just buying more memory. They're trying to secure guaranteed access to avoid a bottleneck they can't afford to leave to chance.
Why this matters for finance professionals
Semiconductor memory now sits at the center of AI infrastructure spending, and Micron's results show that tight supply is translating into multi-year contracted revenue. For investors and analysts, the shift from spot-market pricing to locked-in commitments changes how these stocks should be valued. Finance professionals tracking AI-driven memory demand can benefit from AI for Finance training that connects technology shifts to investment analysis.
Your membership also unlocks:
