Micron quadruples revenue: AI memory pushes margins to 81%

What happened?

Micron Technology, one of the leading manufacturers of DRAM and NAND memory, reported its results for the third fiscal quarter of 2026 (ending in May). The company posted revenue of approximately $42 billion, representing a 367% increase compared to the same period last year, when it generated just over $9 billion. Earnings per share exceeded analyst estimates, and gross margin hit an all-time record of 81.4%, driven by strong demand for high-bandwidth memory (HBM) used in artificial intelligence accelerators.

Why is this important?

This result confirms that Micron has become the primary beneficiary of the generative AI boom, even surpassing other semiconductor manufacturers. Demand for HBM, a type of vertically stacked memory that offers exceptional bandwidth, has surged with the massive adoption of GPUs for training and running AI models. Companies like NVIDIA, AMD, and Intel rely on this memory for their accelerators. Micron's growth is a leading indicator of investment in AI infrastructure and suggests that the data center spending cycle shows no signs of slowing down.

Consequences for the market and industry

Micron's performance has several implications:

• For the supply chain: High demand for HBM is reshaping DRAM manufacturing priorities, with potential bottlenecks in conventional memory production.
• For competitors: Samsung and SK Hynix are also ramping up HBM production, but Micron appears to have gained an edge with its fourth-generation stacking technology (HBM4).
• For prices: Record margins indicate Micron has pricing power, which could lead to more expensive memory for servers and eventually for consumer devices.
• For investors: Micron's stock has risen about 700% over the past two years, and these results could justify higher valuations if AI demand holds.

What should readers know?

Micron's results not only reflect a single company but serve as a thermometer for the AI economy. HBM memory is a critical and scarce component; its availability limits the production of AI GPUs. Additionally, Micron's success could spur new investments in semiconductor manufacturing, especially in the United States, where the company has plants in Idaho and Virginia. However, there is a risk that the upcycle could moderate if hyperscalers (Amazon, Google, Microsoft) reduce their orders for AI servers. For now, the outlook is optimistic: Micron guided revenue for the next quarter above consensus estimates.

"Demand for HBM is unprecedented. We are seeing a structural shift in the memory market driven by generative artificial intelligence," said Micron CEO Sanjay Mehrotra during the analyst conference call.

Comparison with past events

Micron's current growth echoes the DRAM boom during the internet expansion in the late 1990s, though the magnitude is larger. Back then, Micron's revenue grew about 50% annually; today it quadruples in a single year. The key difference is that current demand is concentrated in a single product (HBM) and driven by a specific segment (AI), making the cycle more volatile but also deeper.

Speculation and risks

Not everything is confirmed: some analysts warn that 81% margins may not be sustainable long-term, as competition increases and customers push for lower prices. Moreover, reliance on a single memory type exposes Micron to shifts in AI demand. If large language models hit a performance ceiling and hyperscalers cut spending, revenue could drop sharply. For now, there are no signs of a slowdown.