Micron Technology posted a 15.7% single-day gain on the back of what appears to be a strong earnings catalyst tied to its AI memory business, with FY2025 revenues of $37.4B representing 48.9% year-over-year growth — a remarkable reacceleration for a company that was in a severe cyclical trough just 18 months ago. Gross margins clocked in at 39.8% and diluted EPS hit $7.59, signaling that the pricing environment in DRAM and NAND has shifted materially in Micron's favor, driven largely by insatiable HBM (High Bandwidth Memory) demand from AI accelerator makers.

The move matters because Micron is the primary US-listed pure-play on the memory upgrade cycle underpinning AI infrastructure — its HBM3E is shipping to hyperscaler and GPU partners, putting it directly in the demand stream of Nvidia's next-gen GPU platforms. The juxtaposition with a mixed broader market, where Big Tech AI capex concerns weighed on sentiment, is notable: the market is rewarding the picks-and-shovels supplier even as it questions whether the end-demand from hyperscalers will hold.

The bull case rests on Micron's revenue trajectory — 48.9% YoY growth with expanding margins suggests the memory upcycle still has legs, and HBM supply remains tight well into 2026 based on current industry capacity buildouts. If AI capex from Microsoft, Meta, and Google holds at current levels, Micron's order book stays full and ASP pressure remains subdued.

The bear case centers on the very concern that weighed on the broader market that same session: if Big Tech pulls back on AI infrastructure spending — whether due to ROI skepticism, regulatory pressure, or macro deterioration — DRAM and NAND pricing could reverse sharply, as memory markets are historically the fastest to reprice on demand signals. Memory is a commodity, and a 15.7% single-day gap leaves limited margin of error for any disappointment in forward guidance.

Watch for next-quarter guidance details on HBM shipment volumes and ASP trends, any signals from Nvidia's supply chain commentary, and whether the broader AI capex conversation resolves bullishly or becomes a headwind for the whole infrastructure stack.