Micron lands $22 billion AI supply deals as revenue surges
Micron locked in $22 billion of memory-chip demand as quarterly revenue hit $41.46 billion. Investors are asking whether AI memory, not AI compute, can drive the next winner.
Micron Technology said customers have committed $22 billion to secure memory-chip supply across 16 strategic agreements, a sign that AI demand is turning one of the industry’s most cyclical businesses into a longer-term bet. The Boise, Idaho company said fiscal third-quarter revenue reached $41.46 billion for the quarter ended May 28, up from $23.86 billion in the prior quarter and $9.30 billion a year earlier, and it forecast about $50 billion for fiscal fourth quarter.
The market rewarded the numbers quickly. Micron shares jumped 12% in after-hours trading, a move that reflected how sharply sentiment has shifted around a company that started in 1978 as a four-person semiconductor design shop in the basement of a Boise dental office. Micron described the agreements as “transformational,” and said the contracts include take-or-pay commitments, cash deposits and pricing floors meant to protect margins in a business long defined by boom-and-bust pricing.
The bull case rests on high-bandwidth memory, or HBM, the specialized chip that feeds AI accelerators with data fast enough to keep Nvidia’s systems running. Micron is now the only U.S.-based HBM supplier, competing with Samsung Electronics and SK hynix from South Korea, and its HBM3e products are already in Nvidia’s GB200 systems. HBM4 is slated for production in 2026 for Nvidia’s Rubin platform, linking Micron directly to the next wave of AI infrastructure rather than just today’s buildout. Jensen Huang has repeatedly highlighted the importance of advanced memory to AI, and the supply chain around Nvidia’s next-generation Vera Rubin program has already elevated HBM from a niche component to a strategic bottleneck.
That bottleneck helps explain why Wall Street has chased Micron so aggressively. By March, the company’s stock had tripled in 2025 and risen another 62% year to date, a run that reflected tight supply and surging demand from AI customers. Analysts have compared Micron’s profitability surge to Nvidia’s earlier margin expansion, but the comparison has limits. Nvidia sells the compute that powers AI models; Micron sells the memory that feeds it. The business models are not the same, and Micron still depends on capital spending from cloud and AI customers, while NAND remains weaker than HBM.
Micron’s latest results show why investors are treating memory as a critical layer of the AI stack, but they also show how much of the story still depends on pricing, supply discipline and a narrow group of buyers. The company’s commitments are large, but so is the concentration risk inside a market still shaped by geopolitical rivalry, manufacturing concentration and the familiar volatility of memory chips.
This article was produced by Prism’s automated news system from verified source data, official records, and press releases, then run through automated quality and moderation checks before publishing. The system is built and supervised by the people who set the standards it runs under. Read our full AI policy.
Did this article answer your question?
