MU Slides Another 5% as the Memory-Complex Rerating Drags Into a Second Day

The Second Leg Down

MU's 5.07% slide to $971.80 isn't a fresh shock — it's the continuation of a two-day unwind that started when Broadcom guided Q3 AI chip revenue to $16 billion against a $17.2 billion estimate and, just as damaging, declined to raise its 2026 AI target. Broadcom itself fell more than 12% on the print, and the read-through flushed the entire AI-semis complex. Micron, the most leveraged liquid name to memory pricing, opened down more than 7% on June 4 and has bled further since. The Motley Fool framed it bluntly: this was a sector recalibration, not a Micron event. Nothing in the company's order book changed between Wednesday's close and now.

The Memory-Peak Argument

The Broadcom rout landed on top of a note from Raymond James analyst Karl Ackerman, who argued that DRAM and NAND average selling prices will peak in mid-2026 — roughly a year earlier than the mid-2027 timeline most of the Street was carrying. He sees sequential quarterly price declines starting early next year, driven by rising supply and softer non-AI demand. The nuance the tape ignored: Ackerman kept an Outperform rating, arguing long-term supply agreements should blunt the usual boom-bust whipsaw. In other words, the bear case here is a *flatter* peak, not a cliff — but a memory stock priced for acceleration doesn't trade on nuance when the word 'peak' is in the headline.

Valuation, Not Fundamentals

This is what a positioning flush on a parabolic chart looks like. MU is still up roughly 865% on the year and trades near a 50x trailing P/E, which leaves zero room for a guidance miss or a softening pricing narrative. When a stock is priced for memory ASPs to keep ripping, any analyst willing to put a date on the peak becomes a catalyst by itself. The fundamentals haven't cracked — revenue growth and HBM demand into 2027 are intact — but the multiple was always the fragile part of this trade, and that's exactly where the two-day damage is concentrated.

What June 24 Settles

The argument over whether memory pricing peaks in mid-2026 or mid-2027 can't be resolved by an analyst note — it gets resolved by Micron's own guidance. Fiscal Q3 earnings land June 24, and that print is the first hard read on whether DRAM and NAND pricing is still accelerating or already rolling. Until then, both sides are trading a forecast. A bull who thinks the supply-agreement cushion is real treats this drawdown as a multiple reset to add into; a bear treats it as the first innings of cycle-top de-rating. The $248M in 24h perp volume on the HIP-3 market says traders are taking both sides aggressively into the event.