June 24, 2026
Shortage Turns Into Pricing Power
Micron Technology delivered a forecast that reset expectations for the memory cycle: about $50 billion of revenue for the fiscal fourth quarter ending in August, versus the $43.2 billion analysts expected. Excluding certain items, the company guided to roughly $31 of profit per share, also above the $25.31 projection. The stock rose about 9% in late trading after the announcement, extending a run that had already more than tripled this year. That kind of move shows how quickly a commodity-like semiconductor segment can become a scarcity story when artificial intelligence demand absorbs supply.
The core tension is simple: data center operators need more memory for AI infrastructure, and the industry cannot add enough capacity quickly. Micron, Samsung Electronics and SK Hynix are all benefiting from the surge, especially in high-bandwidth memory, or HBM, which is used alongside AI processors. At the same time, demand for conventional memory has tightened supply for computers, phones and cars. The market is effectively repricing memory from a cyclical component into a strategic bottleneck, at least for now.
For the fiscal third quarter ended May 28, Micron reported $41.5 billion of sales, well above the $35.7 billion analysts expected. Earnings reached $25.11 per share, compared with estimates of $20.49. The margin story was even more striking: adjusted gross margin more than doubled to 84.9%, above the 81.9% estimate. In practical terms, shortages are not just lifting revenue; they are flowing directly into profitability.
Nvidia Pulls the Chain
Micron’s position is tied closely to the AI infrastructure buildout led by Nvidia. The company works with Nvidia to integrate memory into AI systems, and Nvidia Chief Executive Officer Jensen Huang recently confirmed that Micron’s HBM4, along with supply from rivals, will be used for the next-generation Vera Rubin platform. For memory makers, being qualified for these platforms matters because it can determine who captures the most valuable part of the next AI hardware cycle.
SK Hynix currently leads the HBM market and has announced plans for a US stock listing, seeking roughly $29 billion in the offering. That planned capital raise underscores how aggressively memory companies are trying to fund expansion while demand is still running ahead of supply. The risk is that semiconductor cycles can turn, but the current numbers suggest the upcycle has more force than analysts had modeled.
For investment banking candidates, this is a useful case study for interviews because it links operating drivers to valuation: demand growth, capacity constraints, pricing, margin expansion and capital markets activity all appear in one story. A strong answer would not stop at saying AI is good for chips. It would explain why HBM scarcity changes the earnings profile, why Nvidia platform qualification matters, and why a peer such as SK Hynix might choose a large US listing while investor appetite is intense.
The broader takeaway is that AI demand is pushing pressure through the entire hardware supply chain. If memory remains scarce, higher prices may continue affecting not only data centers but also consumer devices and autos. Micron’s forecast is therefore more than a company beat; it is a signal that the AI buildout is still pulling supply, pricing and capital toward the memory sector.