Markets MU JUN 27, 2026

Micron's $41.5B Q3 reverses global AI selloff as Q4 guide hits $50B

MU jumped 15% after-hours on a 346% revenue jump, 84.9% gross margins, and $22B in locked customer cash commitments — unwinding the June 23-24 AI capex selloff.

BY HENLEY MARRAST · MARKETS DESK

Micron reported fiscal Q3 revenue of $41.46 billion against an LSEG consensus of $35.84 billion, sent the stock up 15% in extended trading, and effectively ended the 48-hour AI-capex selloff that had dragged the Nasdaq down 2.2% across June 23-24 and bled into the KOSPI overnight. Revenue grew 346% from $9.3 billion a year earlier. The Q4 guide came in at a record $50.0 billion plus or minus $1.0 billion.

The number that stopped the selloff wasn’t revenue. It was the gross margin: 84.9% adjusted, against a Bloomberg consensus of 81.83%, and 39% in the year-ago quarter. Q4 guidance sets the figure at roughly 86.0%. Memory pricing, which has historically been the most cyclical line in semiconductors, is behaving like a subscription business.

Net income hit $28.24 billion, or $24.46 per share, against $1.89 billion a year ago. Data center revenue alone exceeded $25 billion, putting that segment on a $100 billion-plus annualized run rate. Data center SSD revenue topped $5 billion, more than doubling sequentially. HBM4 12-high is ramping at twice the speed of HBM3E 12-high, with over $1 billion already shipped.

Then the disclosure that reframed the cycle: 16 strategic customer agreements representing $22 billion in financial commitments, with roughly half or more of company revenue now sitting under contract. Anthropic is among the named counterparties. This is what the buy-side wanted evidence of, hyperscaler capex turning into locked memory dollars rather than vibes.

CFO Mark Murphy and CEO Sanjay Mehrotra raised FY26 capex to $27 billion from $25 billion and projected $45 billion for FY27, with Idaho Fab 1 targeting first wafers by mid-2027. JPMorgan analysts Tarek Hamid and Nathaniel Rosenbaum lifted their cumulative AI capex projection through 2030 to $5.5 trillion from $5.1 trillion, with hyperscaler spend hitting $1.1 trillion in 2027 alone. Bank of America’s Vivek Arya pegs memory’s share at roughly 35%.

Operating cash flow was $25.4 billion, free cash flow a record $18.3 billion. Cash closed at $30.2 billion against $5.7 billion of debt, after $4.4 billion was retired in the period.

The 2017-2018 DRAM cycle ended in a glut because demand was consumer-driven and inventories cleared in months. This one is being underwritten by contracted hyperscaler cash, which is a different financial object entirely.

Sources

About the author

Henley Marrast

MARKETS DESK

Henley Marrast covers AI-equity flow, accelerator demand, and earnings prints for AI Sheet Report. She leads coverage of the public AI complex from the New York markets desk, with a focus on the daily tape and quarterly results. She has been writing about technology markets for several years.