Micron's HBM production capacity is expected to double by the end of the year, catching up with Samsung and SK Hynix.

Micron Technology plans to increase its monthly HBM capacity to around 100,000 wafers by the end of this year, nearly doubling last year's output, in a bid to capture a larger share of the global AI memory market.

Article Author, Source: Yonhap News Agency

Micron Technology is aggressively expanding its High Bandwidth Memory (HBM) production capacity, aiming for a larger share of the global AI memory market.

According to industry insiders cited by Yonhap News Agency on the 4th,Micron plans to boost its monthly HBM capacity to about 100,000 wafers by the end of this year, nearly doubling last year's output. By then, the capacity gap between Micron and Samsung Electronics and SK Hynix is expected to shrink to about half of what it is now. At the same time, the company is accelerating the ramp-up to mass production of its latest generation HBM4 12-layer product to meet the needs of NVIDIA's next-generation AI accelerators.

This move reflects continued strong demand in the AI memory market. AI chip makers such as NVIDIA have maintained a high demand for HBM, with supply still unable to keep pace with demand. For Micron, which holds just an 18% market share and has long ranked third, this capacity expansion is not only a key battle for market share, but also an essential step to convert the AI memory demand boom into actual revenue.

Capacity target: Striving for 100,000 wafers per month by year-end

According to industry sources,Micron plans to add up to 60,000 wafers per month to its HBM capacity by the end of the year. Last year, the company's HBM monthly output was about 40,000 to 50,000 wafers, which means the new capacity will surpass the existing base and push total capacity to around 100,000 wafers per month.

To achieve this goal, Micron is increasing purchase orders with multiple HBM manufacturing equipment suppliers, with related equipment continuously shipped to its Taiwan and Singapore factories. Currently, Micron's HBM packaging business is mainly carried out at its Tongluo plant in Taiwan and the Woodlands plant in Singapore, while an equipment investment plan for its Hiroshima, Japan factory is also underway.

In terms of product structure, Micron is rapidly shifting toward HBM4 12-layer.Currently, its HBM output is mainly focused on HBM3E 12-layer, but reportedly the share of HBM4 12-layer will increase from 20%–30% at the beginning of the year to up to 50% by year-end.

HBM4 12-layer is the memory matched for NVIDIA’s latest AI accelerator “Vera Rubin.” Micron began mass production of this product in the second quarter of this year. At the fiscal Q3 2026 earnings call in June, Micron CEO Sanjay Mehrotra stated that HBM4 12-layer's mass production ramp-up is about twice as fast as HBM3E 12-layer, and the cumulative HBM4 shipment revenue has exceeded $1 billion.

Market landscape: The gap between the top three may significantly narrow

Although Micron is one of the world’s top three HBM manufacturers, there is still a clear capacity gap compared to Samsung Electronics and SK Hynix. Industry consensus holds that both Samsung and SK Hynix have a monthly HBM output of about 150,000 to 200,000 wafers each—3 to 4 times that of Micron. Unless Micron narrows this gap, it will be difficult to change its long-standing third-place position in the market.

According to market research firm Counterpoint Research, SK Hynix held a 50% share of the global HBM market in the second quarter of this year, followed by Samsung Electronics at 32% and Micron at 18%. If the capacity expansion target is achieved as planned, the capacity gap between Micron and its two main rivals could be cut to half of its current level, possibly reshuffling the market share distribution.