Micron Posts $54.23 Billion Quarter as AI Memory Demand Explodes

Micron reported adjusted earnings of $33.42 per share on revenue of $54.23 billion for its fiscal fourth quarter, beating Wall Street expectations on both lines as demand for AI memory shows no sign of cooling. The stock rose slightly in extended trading on Wednesday.

Analysts polled by LSEG had expected $31.61 in adjusted earnings per share on $51.07 billion in revenue. Micron delivered nearly $3.2 billion more in sales than consensus and beat on the bottom line by more than a dollar and a half per share.

The quarter's growth rate is the story. Revenue almost quadrupled from $11.32 billion in the year-earlier quarter, according to the company's statement. Net income climbed to $37.7 billion, or $32.87 per share, from $3.2 billion, or $2.83 per share, a year ago — a nearly twelvefold increase in profit.

Guidance points higher

Micron's outlook for the fiscal first quarter came in above what analysts anticipated. The company said it expects revenue of about $61.5 billion and adjusted earnings per share of $38.15. Analysts polled by LSEG had expected $35.40 in adjusted earnings per share on $57 billion of revenue.

That guidance implies continued sequential growth and another significant beat against consensus if Micron delivers. It also signals that the memory maker sees the current demand environment extending at least through the coming quarter.

One portfolio manager who follows the company closely sees the cycle running much longer than that. Hendi Susanto, portfolio manager at Gabelli Funds, called the results "another strong beat and raise for Micron" in an email after the release. He added: "At this point, I have not heard any negative data points pointing to the memory cycle reversing toward a decline anytime soon for the foreseeable future."

DRAM dominates, HBM drives it

Micron is the only U.S.-based maker of high-bandwidth memory, the advanced memory technology built from stacks of general-purpose DRAM that has become a critical component of AI accelerators. Advanced graphics and central processors from chip giants like Nvidia and AMD need increasing amounts of HBM to handle AI workloads, and the world's leading providers cannot make enough.

That shortage is visible in Micron's segment results. Fourth-quarter DRAM revenue increased 343% from a year ago to $39.8 billion, representing 73% of total sales.

The supply crunch has broader consequences beyond Micron's income statement. The shortage has led to a spike in memory costs and resulted in increased prices for consumer electronics like Apple's iPads and MacBooks. In other words, the AI buildout is now showing up in the price tags of hardware that has nothing to do with training or running models — a concrete example of how data center demand propagates through the broader electronics supply chain.

A $250 billion manufacturing bet

To capitalize on the demand, Micron is investing $250 billion to build two new campuses for making HBM. The largest broke ground in Clay, New York, in January. Its first new fab in Boise, Idaho, is scheduled to come online next year.

Micron's hometown buildout in Boise alone totals $50 billion.

The company is racing rivals who are scaling just as aggressively. HBM leaders SK Hynix and Samsung are in the midst of massive new HBM factory buildouts in their home country of South Korea. Micron holds the smallest HBM market share of the three, but its market capitalization has still topped $1.2 trillion — a figure that reflects how much investors have repriced the memory business in the AI era.

The stock has soared more than 500% in the past year, driven by a worldwide supply crunch caused by what the company describes as historic levels of demand for memory chips needed for AI models and workloads.

Policy proximity

Micron's results arrive at a moment when AI infrastructure sits at the center of both industrial policy and geopolitics. CEO Sanjay Mehrotra participated in a summit on AI regulation hosted by President Donald Trump on Tuesday, days after attending a White House dinner with Chinese President Xi Jinping on Xi's first visit to the U.S. in more than a decade.

The appearances place the head of the only American HBM maker directly in the room as U.S. leadership shapes the regulatory and trade environment around the technology driving Micron's growth. For a company whose product sits on the critical path of every major AI accelerator, regulatory posture toward the industry is not an abstract concern.

Why it matters

Micron's quarter is one of the cleanest readouts available on the physical economics of the AI boom. Hyperscalers and chip designers can talk about demand; Micron sells them the memory they cannot get enough of, and its near-quadrupling of revenue quantifies the gap between AI infrastructure appetite and global manufacturing capacity.

The stakes run in two directions. If the cycle holds, Micron's $250 billion campus investment and its fiscal first-quarter guidance of $61.5 billion in revenue mark the early innings of a sustained expansion, with DRAM pricing strength continuing to flow through to earnings. If memory costs keep climbing, consumers will keep paying for it in higher prices for devices like iPads and MacBooks.

For now, the data points all point one way. As Gabelli's Susanto put it, he has not heard anything suggesting the memory cycle reverses toward decline "anytime soon for the foreseeable future." The first test of that view arrives when Micron reports its fiscal first quarter and shows whether the $61.5 billion revenue target — and the capacity coming online in Boise next year — can keep pace with demand that the industry still cannot fully supply.