Micron stock surges 16% on blockbuster earnings; revenue 4.4x YoY amid AI memory boom
Micron stock soared 16.4% in premarket trading on Thursday after reporting blockbuster third-quarter earnings driven by surging demand for memory in AI infrastructure. The company's revenue more than quadrupled from $9.3 billion a year earlier to $41.46 billion in fiscal Q3, beating analyst expectations of $36 billion. Micron is now forecasting revenue of approximately $50 billion for the current quarter, up from $11.3 billion in the prior year. The stock rose 723% over the past year, pushing the company's market cap to $1.2 trillion.
Micron has benefited from the AI infrastructure buildout by hyperscalers—AI data centers require massive amounts of memory chips, reducing supply for smartphones and PCs and pushing memory prices higher. The company signed 16 long-term supply agreements with customers spanning data centers to automakers, locking in sales for three to five years with $22 billion in financial commitments. The company expects about 40% of its revenue to come from long-term contracts with minimum prices built in, providing downside protection even if demand weakens.
The earnings have lifted global tech stocks. In premarket trading, Qualcomm was up 12%, Intel +6%, AMD +3.6%, and NVIDIA +1.5%. RBC Capital Markets analysts said their base case is for the current upcycle to continue through 2027, and the supply agreements give conviction on sustainability. Analysts raised estimates, price targets, and reiterated Outperform ratings.
For infrastructure teams and investors, Micron's results validate the AI capex cycle thesis: memory supply is the binding constraint on data center expansion, and long-term pricing power is real. The 40% locked-in contract revenue with minimum prices means Micron has hedged margin risk even in a demand slowdown. This shifts competitive pressure to spot market pricing for non-contracted capacity and raises questions about inventory management as the cycle matures through 2027.
Sources
- Primary source
- cnbc.com
“AI data centers require large amounts of memory chips”