Micron beat Wall Street estimates with fiscal fourth-quarter revenue of $54.23 billion and told investors the memory shortage will intensify through fiscal 2027 and 2028, guiding the current quarter to $61.5 billion in revenue as data center orders crowd out consumers. Shares recovered from an initial dip and rose roughly 2 percent in after-hours trading on a report that beat both the top and bottom line.
The tally against analyst consensus, compiled by LSEG, was decisive. Adjusted earnings per share came in at $33.42 against $31.61 expected, and revenue of $54.23 billion topped the $51.07 billion estimate. GAAP net income reached $37.70 billion, or $32.87 per diluted share. Fiscal Q4 revenue climbed 31 percent from the prior quarter and 379 percent from the year-ago quarter, marking the company’s sixth consecutive quarterly revenue record. For the full fiscal year, Micron posted $133.19 billion in revenue, up 256 percent, with gross margin for the year at 80.7 percent, nearly double the prior year’s 39.8 percent. Non-GAAP net income for the year reached $86.76 billion, and management said gross margin, revenue and earnings per share all exceeded the high end of the company’s own guidance for the quarter. Operating cash flow for the fiscal year came in at roughly $51 billion, cash that funds the company’s next wave of factory investment, and gross margin for the quarter itself hit 86.8 percent non-GAAP, above what the market had been modeling.
The demand engine behind those numbers is the same one driving the rest of the AI buildout. Micron’s Cloud Memory Business Unit, which sells high-bandwidth memory and data center DRAM into AI server programs, generated a record $16.3 billion in the quarter, about 30 percent of total company revenue. Management said on the earnings call that industry demand has strengthened since the last update and that memory and storage supply-demand conditions will be much tighter in fiscal 2027 and 2028 than they were in fiscal 2026, a statement that reads as a pricing signal as much as a demand one. The company framed the year as record fiscal 2026 results with AI-driven demand positioning it for a record fiscal 2027.
Shortage spreads to gadgets
The spillover is already visible in consumer prices. CNBC reported that the shortage has pushed up memory costs and, with them, prices for gadgets such as Apple’s iPads and MacBooks, because DRAM and NAND are significant line items in a device bill of materials and contract prices have been climbing for months. Memory makers are allocating supply toward high-margin data center contracts, leaving consumer electronics makers to bid for what remains, and some device vendors have started passing the difference through to shoppers ahead of the holiday quarter. Micron’s fiscal Q1 2027 guidance, revenue of $61.5 billion plus or minus $1.5 billion with diluted earnings per share of $38.15 plus or minus $1.00 on a non-GAAP basis, assumes those conditions hold into the December quarter.
The board declared a quarterly dividend of $0.15 per share at the results, payable October 29 to shareholders of record on October 14. Micron’s fiscal year ended September 3, and the company announced the record full-year results in a press release dated September 30, the same day the Nasdaq closed out its strongest quarter since 2020 on AI-related optimism and Nvidia shares pushed further into record territory. A year earlier the same quarter produced $11.32 billion of revenue, so the scale of the shift is easier to grasp as a multiple than as a percentage. In plain terms, the average AI server now carries several times the DRAM of the general-purpose machines it replaces, which is why a supplier with the right HBM lines can post numbers the rest of the industry cannot approach.
Where the bottleneck sits
On the call, management pinpointed the constraint less at wafer fabrication and more at packaging and test capacity for advanced stacks, which is where most new capital spending is being directed. High-bandwidth memory requires stacking DRAM dies and testing them at yields the industry is still ramping, and every additional HBM stack consumes more wafer capacity than a conventional module, which tightens the overall supply picture even when raw fab output grows. Management also noted that data center customers might despec products, using lower-grade memory when shortages bite, a step that would have been unthinkable in a normal supply year and a sign of how stretched allocations have become.
What comes next
For the broader hardware market, the report lands in the middle of a pricing cycle that analysts including Morgan Stanley and KeyBanc expect to stretch across multiple quarters, with memory makers benefiting alongside foundry and advanced packaging providers. Samsung and SK Hynix both supply HBM into the same Nvidia programs, and industry trackers expect all three to allocate more wafer capacity to data center memory through 2027, keeping spot prices elevated even if unit shipments grow. For consumers, the practical effect shows up in upgrade cycles and new device pricing over the coming months.
For Micron, the results cap a fiscal year in which the AI data center buildout turned a cyclical commodity business into one of the S and P 500’s fastest growers, and management’s own framing was blunt: they expect fiscal 2027 to be even better. Sanjay Mehrotra, Micron’s chief executive, credited employees around the world in a closing statement on the call, thanking them for what he described as above and beyond efforts that made the results possible. Wall Street’s read will show up in the stock’s next session, but the guidance has already moved the bar upward once again.
Sources here for reference. Micron press release of September 30, 2026; CNBC; LSEG consensus figures; the Micron fiscal Q4 2026 earnings call transcript.