Micron breaks ground on Singapore fab as it commits about $24 billion to expand NAND capacity
Micron is expanding its Singapore manufacturing footprint with an advanced wafer fabrication facility and a roughly $24 billion, 10-year investment plan, aiming to meet AI-driven demand for memory and storage and to strengthen supply resilience.

A decade-long expansion centered on AI-era demand
Micron said it has broken ground on an advanced wafer fabrication facility in Singapore, describing the project as a planned investment of roughly $24 billion over 10 years. The company positioned the new fab as a response to growing long-term demand for NAND technology, which is being pulled higher by the expansion of AI workloads and other data-centric applications. Micron described the site as part of its existing NAND manufacturing complex, effectively expanding a location that is already central to its global flash-memory output.

According to the company, the facility is designed to ultimately provide about 700,000 square feet of cleanroom space, with wafer output scheduled to begin in the second half of calendar 2028. Micron also framed the investment as a way to improve supply-chain resilience by deepening manufacturing capacity in a key hub that already hosts both production and related R&D activity.
Singapore’s role in the memory supply chain
Micron highlighted Singapore as a long-standing manufacturing base and said the groundbreaking ceremony included senior Singapore government and agency officials. The company cast the project as both an industrial-scale capacity build and a platform for faster technology transitions, with manufacturing and research co-located to reduce friction between development cycles and high-volume production.
The expansion lands at a time when memory markets are heavily influenced by AI infrastructure buildouts and by attempts to avoid the boom-bust oversupply cycles that have historically hit DRAM and NAND. By phasing production for a 2028 start and describing cautious scaling, Micron is signaling that it wants more capacity—without recreating the conditions that previously collapsed pricing across the memory segment.