Micron Posts Record Margins, $100B in AI Memory Contracts

Micron Technology delivered record financial results for its fiscal third quarter and disclosed it has locked in approximately $100 billion in contracted memory revenue through strategic customer agreements, signaling that AI infrastructure builders view high-bandwidth memory as a critical constraint worth securing years in advance.

The announcement came one day after semiconductor stocks suffered their second-worst session of the past year, with Micron shares hit particularly hard amid broader concerns about AI chip demand.

Record profitability amid AI buildout

Micron reported revenue of $41.5 billion for the quarter, surpassing Wall Street expectations. Adjusted earnings reached $25.11 per share. The standout metric: gross margin hit 84.9%, more than double the year-ago figure and the highest level in company records dating to 1990.

Gross margin measures how much revenue remains after manufacturing costs. At nearly 85%, Micron is retaining an unprecedented share of each sales dollar, a sign that pricing power remains strong despite recent market jitters about AI demand sustainability.

The company projected gross margins will climb further to approximately 86% in the current quarter, contradicting concerns that the memory market is softening.

Customers lock in supply with binding contracts

Micron disclosed it has signed 16 strategic customer agreements designed to guarantee memory supply over multiple years. Fourteen of those deals represent roughly $100 billion in minimum contracted revenue over their remaining terms, backed by $22 billion in cash deposits and related commitments.

The agreements include take-or-pay provisions, meaning customers must either purchase the contracted volumes or pay penalties. These are not typical purchase orders but binding commitments that span several years.

For a memory industry historically defined by boom-and-bust cycles driven by spot pricing, the shift toward long-term contracted supply represents a structural change. AI system builders appear willing to sacrifice pricing flexibility to ensure they won't face memory shortages that could bottleneck their infrastructure deployments.

Why it matters

The contracted revenue and record margins suggest AI infrastructure demand is creating a fundamentally different supply dynamic for high-bandwidth memory than existed in previous chip cycles. Hyperscalers and AI hardware makers are treating memory capacity as a strategic asset rather than a commodity input, locking in supply years ahead to avoid potential bottlenecks. For investors weighing concerns about AI spending sustainability, Micron's results indicate customers are committing capital at scale with multi-year visibility.

Memory as AI infrastructure constraint

Micron emphasized that AI accelerators require memory that is physically proximate, extremely fast, and available in massive quantities. The company positioned memory performance and capacity as direct determinants of overall AI system performance, elevating memory from a supporting component to a first-order constraint.

The financial results and contract disclosures were first reported by Yahoo Finance.