Micron posts record $41.5bn Q3 revenue on AI memory surge

Micron's fiscal Q3 2026 revenue rose 346% year-on-year, driven by AI-related demand for HBM and data-centre memory products.

Micron picks Bechtel to build first phase of New York mega-fab

Micron Technology has reported record quarterly revenue of $41.46 billion for its fiscal third quarter ended 28 May 2026, a 346% increase on the $9.30 billion posted in the same period a year earlier and a 74% jump on the prior quarter's $23.86 billion. GAAP net income reached $28.24 billion, or $24.67 per diluted share, while adjusted free cash flow hit $18.3 billion. The Boise, Idaho-based chipmaker ended the quarter with $30.2 billion in cash, marketable investments and restricted cash.

The results reflect surging enterprise demand for high-bandwidth memory and data-centre DRAM, with the company's two largest business units posting exceptional growth. The Cloud Memory Business Unit generated $13.77 billion in revenue at an 83% gross margin, while the Core Data Centre Business Unit contributed $11.52 billion at 87% gross margin. The Automotive and Embedded Business Unit also expanded sharply, reaching $4.63 billion from $1.13 billion a year earlier, reflecting growing memory content in electric vehicles and advanced driver-assistance systems.

Product roadmap

On the product side, Micron confirmed that HBM4, built on its 1-beta DRAM technology, is in high-volume shipments for a lead customer and that qualification samples have reached multiple additional end-customers. Development of HBM4E, using next-generation 1-gamma DRAM technology, is under way with volume production targeted for calendar 2027. The company also shipped qualification samples of 256GB DDR5 RDIMMs built with 1-gamma technology and advanced 3D die stacking to server ecosystem partners, and commenced shipments of its high-capacity 245TB QLC solid-state drive.

Sanjay Mehrotra, chairman, president and chief executive, said: "Micron is investing at record levels in technology, products and supply to address our customers' rapidly growing demand. We believe our multi-year Strategic Customer Agreements will significantly enhance the durability and predictability of Micron's strong financial performance."

Capital expenditure for the quarter was $7.1 billion net, and Micron's board declared a quarterly dividend of $0.15 per share, payable on 21 July 2026. For the fourth quarter, Micron guided for revenue of approximately $50 billion, a further sequential step-up, with gross margin expected to reach approximately 86% and diluted earnings per share of around $31.00 on a non-GAAP basis.

Market context and competitive landscape

Micron's trajectory sits within a broader AI infrastructure investment cycle that is reshaping the memory and storage industry. Demand for HBM in particular is being driven by the proliferation of large-scale GPU clusters in hyperscaler and sovereign AI deployments, where memory bandwidth is a primary bottleneck. SK Hynix has been the dominant HBM supplier to date, with Samsung working to qualify its own HBM3E at scale. Micron's rapid ramp of HBM4 positions it as a credible third supplier, and the company's reference to multi-year Strategic Customer Agreements suggests it has secured committed long-range purchasing commitments, though it has not publicly named the counterparties.

The scale of Micron's revenue growth also raises questions about cyclicality. The memory industry is historically prone to sharp boom-and-bust cycles driven by oversupply; the current upcycle is unusually demand-led rather than capacity-push, which may make it more durable. That said, Micron's $19.6 billion in property, plant and equipment expenditure across the first nine months of fiscal 2026 represents a substantial supply commitment that would weigh heavily on margins if AI server build-out slows or shifts to alternative architectures. Regulators in the US and EU are also closely watching semiconductor supply chains: US CHIPS Act funding, of which Micron is a recipient, carries domestic-production conditions that constrain where leading-edge capacity can be built.