Micron bets big on AI-era memory
Micron Technology, the largest US memory-chip maker, said it plans to invest $200 billion to expand production capacity as artificial intelligence drives what executives describe as the most severe memory supply crunch in more than four decades.
The investment is aimed at scaling domestic manufacturing for chips that underpin AI servers, smartphones, PCs, vehicles and hyperscale data centres. Demand for high-performance memory—particularly DRAM used in high-bandwidth memory (HBM) stacks—has accelerated faster than many in the industry expected, tightening supply and pushing customers toward longer-term purchasing agreements.
Idaho expansion anchors the plan
At its headquarters in Boise, Idaho, Micron is allocating $50 billion to more than double its 450-acre campus. Two large fabrication plants are under construction. The first, known as ID1, is expected to produce initial silicon wafers in mid-2027 and focus on DRAM for HBM, a key component for advanced AI computing. Both facilities are slated to be operational by the end of 2028.
Company executives say the constraint is not willingness to invest but the time required to add clean-room capacity at scale. Mark Murphy, Micron’s CFO, told investors the company is “doing everything we can to add capacity,” while noting “there is no easy or fast way to get that done.”
From commodity cycles to strategic asset
Memory has traditionally been prone to boom-and-bust pricing. After a pandemic-era surge faded in 2022, the industry cut output to stabilise markets. The AI buildout has reversed that dynamic, with reports of sharp increases in DRAM contract prices and steep rises for newer standards such as DDR5. Market participants say buyers are increasingly seeking multiyear contracts to lock in supply.
Beyond Idaho, Micron has broken ground on a $100 billion fab complex near Syracuse, described as the largest private investment in New York state history, and announced a $9.6 billion expansion in Hiroshima. Rival SK Hynix is also expanding, including new projects in South Korea and Indiana, underscoring the high-stakes race to supply the AI economy.
Related
