Micron is doubling down on artificial intelligence memory production to close the gap on rival chipmakers. A fresh report from Korean outlet ETNews indicates the memory giant will double its monthly high bandwidth memory wafer output to 100,000 units before the year ends. The aggressive surge aims directly at supplying next generation AI hardware like Nvidia Rubin chips.
The core of this factory ramp centers on 12Hi HBM4 memory stacks. These advanced components are built to supply hungry AI accelerators like the Nvidia Rubin processor line. Early this year, 12Hi HBM4 made up just 20% to 30% of total output for Micron. By the final quarter, the company plans for this specific high end stack to represent 50% of its entire memory output.
Meeting these targets requires massive physical upgrades across manufacturing facilities. Micron is spending heavily on specialized fabrication machinery to keep the schedule on track. Equipment deliveries are arriving at fabrication facilities across Taiwan and Singapore. These installations will allow factories to process silicon at double the pace seen last year, climbing from 40,000 to 50,000 monthly wafers up to the target 100,000 mark.
Even with this fast jump, Micron still faces steep competition from the market leaders. SK Hynix and Samsung currently hold comfortable leads, with both rivals producing between 150,000 and 200,000 wafers per month. But doubling output in a single year shows Micron is ready to fight for a much bigger piece of the AI hardware market.
