Micron Faces Challenges Ahead of Earnings, but Analysts Remain Optimistic

Micron Technology shares fell for a third straight session on Monday, weighed down by worries about competition in the high-bandwidth memory market. Still, several analysts raised their price targets ahead of the company’s March 18 earnings report.

Shares dropped about 1.2% in premarket trading to around $93, continuing a decline that has wiped out about 20% of the stock’s value since its record high in early February. On Friday alone, shares fell 6.7%.

Investors are mainly debating whether demand for memory will last as artificial intelligence infrastructure grows. Analysts at Citi and Susquehanna maintained their positive outlooks and raised their price targets, citing strong data-center spending as a driver of continued pricing strength. Citi increased its target to $430, predicting that DRAM prices could rise by more than 170% in 2026, thanks to large investments from hyperscalers. Their supply chain analysis suggests much of next year’s higher spending will go toward increased memory costs. Susquehanna raised its target sharply to $525, saying supply and demand could balance out by mid-2027 as new factories open.

However, more demanding AI workloads could help protect margins at that point. A more pressing issue is Micron’s position in next-generation high-bandwidth memory. On Sunday, Korean media reported that SK Hynix and Samsung Electronics were chosen as the first exclusive suppliers of HBM4 chips for Nvidia’s upcoming Vera Rubin accelerator platform. If true, this would be a setback for Micron as it tries to catch up with its Korean competitors in the high-end AI memory market.

Independent analyst Richard Windsor tempered the bearish view of the development, suggesting that Micron is likely to be integrated. Independent analyst Richard Windsor offered a more balanced view, saying Micron will probably join Nvidia’s supply chain later as production increases. He also noted that, since capacity is fully booked through 2026, any chips originally meant for Nvidia could easily be sold to other customers without hurting revenue trajectory or visibility into hyperscaler demand beyond the current cycle.