Micron breaks ground on $24 billion Singapore chip fab as AI demand tightens memory supply

Micron Technology said it has broken ground on an advanced wafer fabrication facility in Singapore, a project the company expects will require roughly $24 billion in investment over 10 years. The plant is designed to expand Micron’s long-term manufacturing capacity for NAND flash memory—components used in everything from consumer devices to the storage-heavy infrastructure behind modern AI systems.

Micron’s timing reflects a market reality that has reshaped the semiconductor landscape: AI data centers are consuming vast amounts of memory and storage, tightening supply for other industries and putting upward pressure on prices. Reuters reported that Micron is moving to boost output amid what it described as an acute shortage, as companies across sectors compete for constrained memory supply.

The company said wafer output from the new Singapore facility is expected to begin in the second half of 2028, and the project will ultimately add significant cleanroom space. Micron also indicated it will maintain flexibility in how quickly it ramps capacity, a nod to the memory sector’s history of boom-bust cycles when expansions overshoot demand.

Singapore has long been a core node in Micron’s global manufacturing footprint, and the new project deepens that bet. Company statements emphasized that co-locating research and development with manufacturing can speed technology transitions and improve time-to-market, which is increasingly important as customers demand higher density, better performance, and energy efficiency for AI workloads.

Micron’s announcement also fits into a broader industry scramble. Chipmakers worldwide are adding advanced packaging and new fabs to keep pace with AI infrastructure buildouts, while governments compete to attract investment for strategic supply-chain reasons. In Micron’s case, the Singapore expansion complements other large-scale plans, including U.S. investments intended to diversify production and reduce risk.

For the tech sector, the headline isn’t just the dollar figure—it’s what it signals: memory has become a central bottleneck for AI growth. Faster compute is useless without enough high-performance memory and storage to feed it, and shortages can ripple quickly into cloud pricing, device costs, and the pace at which new AI services can scale.

Investors typically watch two things after announcements like this: whether demand stays strong enough to justify sustained capital spending, and whether capacity additions arrive in time to ease shortages without triggering oversupply. Micron is trying to thread that needle—expand enough to serve AI-era demand, while avoiding the industry’s old habit of building too much too fast.