Micron invests US$24 billion on new Singapore manufacturing facility

The US memory chipmaker's decade-long investment underscores the severity of a supply crunch that analysts predict will persist through late 2027.

US memory chipmaker Micron Technology is doubling down on Singapore with a US$24 billion manufacturing facility, marking one of the largest semiconductor investments in the region as the industry grapples with an unprecedented shortage of memory chips driven by explosive AI demand.

The new advanced wafer fabrication plant, which Micron announced on Tuesday, will take shape over the next decade and begin production in the second half of 2028. The facility will sprawl across 700,000 square feet of cleanroom space dedicated to producing NAND flash memory chips—the storage technology increasingly critical to AI applications and data centers.

The investment comes as tech giants and cloud service providers find themselves locked in a fierce scramble for memory chips of all types. From consumer electronics manufacturers to AI infrastructure providers, the supply crunch has created bottlenecks across the technology sector, with no immediate relief in sight.

“Micron’s leadership in advanced memory and storage is enabling the AI-driven transformation reshaping the global economy,” said Manish Bhatia, executive vice president of global operations at Micron Technology.

“We are grateful for the longstanding support and successful partnership with the Singapore government, including EDB and JTC. This investment underscores Micron’s long-term commitment to Singapore as an important hub in our global manufacturing network, enhancing supply chain resiliency and fostering a vibrant ecosystem for innovation,” he added.

Singapore has emerged as Micron's primary manufacturing hub, with the company producing 98% of its flash memory chips in the city-state. The latest announcement builds on an existing US$7 billion investment in an advanced packaging facility for high-bandwidth memory (HBM)—the specialized chips that power AI accelerators from companies like Nvidia.

That HBM packaging plant remains on track to begin contributing to supply in 2027, Micron confirmed. The memory shortage has sparked an industry-wide capacity race. Micron's South Korean rivals, Samsung and SK Hynix, are advancing production timelines for new facilities in an attempt to capitalize on the supply gap. Earlier this month, SK Hynix told Reuters it would accelerate the opening of one new factory by three months while beginning operations at another plant in February.

Micron is also expanding beyond Singapore. Last week, the company disclosed talks to acquire a fabrication site from Taiwan's Powerchip for US$1.8 billion, a move aimed at boosting its DRAM wafer output. The multi-pronged expansion strategy reflects the urgency chipmakers feel to scale production before competitors capture market share.

Despite these capacity additions, analysts believe the memory shortage will persist well into 2027. The lag between announcing new facilities and actual production—often measured in years—means near-term supply constraints will continue driving prices higher and forcing technology companies to carefully manage their memory chip allocations.

For Micron, which held a 13% share of the flash memory market in the third quarter of 2025 according to TrendForce data, the Singapore investment represents both a response to immediate market pressures and a long-term bet on AI's trajectory. As data-centric applications proliferate and AI models grow more sophisticated, memory requirements are expected to increase exponentially.

The US$24 billion commitment to Singapore also underscores the geopolitical dimensions of semiconductor manufacturing. As tensions between the US and China continue to shape technology supply chains, American chipmakers are consolidating production in allied territories with stable regulatory environments and skilled workforces.

With wafer output not expected until 2028, enterprises and cloud providers will need to navigate at least three more years of tight memory supplies and elevated costs—a reality that may reshape infrastructure investment strategies and slow the pace of AI deployment for some organizations.