Micron HBM Capacity Could Double by Year-End, Closing Gap with Samsung, SK Hynix

Micron Technology plans to significantly boost its monthly HBM capacity to about 100,000 wafers by the end of this year, nearly doubling from last year, to narrow the capacity gap with SK Hynix and Samsung Electronics. The company is accelerating mass production of its 12-layer HBM4 product for Nvidia's "Vera Rubin" architecture, expecting its share of capacity to rise to 50% by year-end, aiming to break out of its current 18% market share position.

Micron Technology is aggressively expanding its high-bandwidth memory (HBM) production capacity, targeting a larger share of the global AI memory market.

According to industry insiders cited by Yonhap News on the 4th, Micron plans to increase its monthly HBM capacity to about 100,000 wafers by the end of this year, nearly double from last year, which could narrow the capacity gap with Samsung Electronics and SK Hynix to about half of the current level. Meanwhile, the company is accelerating the production ramp of its latest 12-layer HBM4 product to meet demand for Nvidia's next-generation AI accelerators.

This move reflects the sustained strong demand in the AI memory market. AI chipmakers like Nvidia have high demand for HBM, and supply still cannot keep up with demand. For Micron, which holds only 18% market share and has long been in third place, this capacity expansion is both a key battle for market position and a necessary step to convert AI memory demand into actual revenue.

Capacity Target: Monthly Capacity to Hit 100,000 Wafers by Year-End

According to industry insiders, Micron plans to add up to 60,000 wafers per month of HBM capacity by the end of this year. Last year, the company's monthly HBM production was about 40,000 to 50,000 wafers, meaning the new capacity will exceed the original base, bringing total capacity to about 100,000 wafers per month.

To achieve this goal, Micron is increasing purchase orders from multiple HBM manufacturing equipment suppliers, with equipment being shipped to its factories in Taiwan and Singapore. Currently, Micron's HBM packaging operations are mainly at its Tongluo plant in Taiwan and Woodlands plant in Singapore, and equipment investment at its Hiroshima plant in Japan is also under preparation.

In terms of product mix, Micron is rapidly shifting toward 12-layer HBM4. Currently, its HBM production is still dominated by 12-layer HBM3E, but the share of 12-layer HBM4 is expected to rise from 20% to 30% at the beginning of this year to up to 50% by year-end.

12-layer HBM4 is the companion memory for Nvidia's latest AI accelerator "Vera Rubin," and Micron began mass production of this product in the second quarter of this year. Micron CEO Sanjay Mehrotra said at the fiscal 2026 third-quarter earnings call in June that the production ramp of 12-layer HBM4 is about twice as fast as that of 12-layer HBM3E, and cumulative HBM4 shipment revenue has exceeded $1 billion.

Market Landscape: Gap Among Top Three May Narrow Significantly

Although Micron is one of the world's top three HBM manufacturers, its capacity still lags significantly behind Samsung Electronics and SK Hynix. Industry consensus is that Samsung and SK Hynix each have monthly HBM capacity of about 150,000 to 200,000 wafers, three to four times that of Micron. If Micron cannot narrow this gap, it will be difficult to change its long-standing third-place market position.

According to data from market research firm Counterpoint Research, in the second quarter of this year, SK Hynix led the global HBM market with a 50% share, Samsung Electronics held 32%, and Micron held 18%. If the capacity expansion target is achieved as planned, the capacity gap between Micron and its two major competitors could be compressed to half of the current level, potentially leading to a redistribution of market share.