SILICON VALLEY — Micron Technology, the U.S. memory chipmaker widely watched as a barometer for the semiconductor market, posted record results for both the quarter and the full year on Sept. 30, as artificial intelligence demand drove an 11-fold surge in data center revenue from a year earlier. The results push back against concerns that the market has peaked and reaffirm the memory upcycle, raising expectations that Samsung Electronics (005930.KS) and SK hynix (000660.KS) will also benefit.
Micron said revenue reached a record $54.23 billion (about 73.704 trillion won) in the fourth quarter of fiscal 2026, which ran from May 29 to Sept. 3. It was the company's sixth consecutive record quarter.
Quarterly revenue nearly quadrupled from $11.32 billion a year earlier, beating the $51.07 billion average estimate compiled by the London Stock Exchange Group. Annual revenue jumped 256% to $133.19 billion from $37.38 billion, also a record.
Adjusted earnings per share came in at $33.42 for the fourth quarter, above the $31.61 consensus estimate. Micron said it expects first-quarter fiscal 2027 revenue of $61.5 billion and adjusted earnings per share of $38.15, exceeding LSEG estimates of $57 billion and $35.40.
The data center business drove the record, with revenue of $18 billion, more than 11 times the $1.577 billion posted a year earlier. The gain reflects sharply higher demand for high-bandwidth memory as AI data centers expand.
Chief Executive Sanjay Mehrotra said he expects even stronger results in fiscal 2027. "We are gaining confidence that Micron's financial performance will become more robust through our strategic customer agreements," he said. Under the SCA structure, customers lock in memory purchase volumes and amounts in advance. The arrangement favors Micron because it blocks price declines that typically occur under conventional long-term contracts.
Micron's SCA count rose to 26 this quarter from 16 in the previous quarter, and the agreements are estimated to account for more than 35% of revenue through 2030. About 75% of the contracts either set floor and ceiling prices or fix prices outright, while the remainder determine prices through additional negotiations based on market rates. As the number of agreements grew, remaining performance obligations climbed 50% to $150 billion from $100 billion in the previous quarter. RPO refers to contracted revenue not yet recognized.
Mehrotra said that after data centers, physical AI applications such as self-driving cars will generate substantial demand starting in 2030. "Autonomous vehicles at the highest level, L4 and above, will carry more than 200 gigabytes of memory and terabyte-scale storage devices such as solid-state drives," he said. "That is 10 times the memory and storage capacity of today's L2 or L3 vehicles, and humanoid robots will generate similar levels of storage demand."
The outlook for HBM, considered the biggest bottleneck in AI, is also upbeat. Micron, which has raised HBM prices for next year, is developing custom HBM4E, its seventh generation, with Nvidia and expects HBM growth to outpace DRAM by 2028. To expand supply of HBM and other memory, the company is building new chip plants in New York state and in Boise, Idaho, where it is headquartered.
Micron is also sharply increasing capital spending to ease supply constraints. First-half fiscal 2027 capital expenditures are projected at $25 billion, including $11.5 billion in the first quarter and $13.5 billion in the second. That means spending close to this year's full-year total of $27.37 billion will be deployed in a single half-year period. Most of the investment will go toward building production facilities.
As a result, investors are watching for margin erosion from the capital spending surge and higher performance bonuses. Micron's gross margin rose to 87% in the fourth quarter from 84.9% in the third, beating the 86.2% consensus. But it is expected to fall to 86.25% next quarter, below the 86.7% forecast. That is why the stock was little changed in early trading despite the record results. Melissa Otto, head of Visible Alpha research at S&P Global, told Bloomberg that investors will want to see how future pricing affects gross margins and the sustainability of earnings.
