Micron Technology is making one of the boldest industrial bets in modern U.S. history. The country’s largest memory-chip manufacturer plans to invest $200 billion to expand domestic production of the silicon chips that quietly power AI systems, smartphones, laptops, vehicles, and vast data centres.
The company is racing to prevent what executives describe as the worst supply crunch the memory industry has faced in more than four decades. Notably, demand is accelerating faster than few anticipated.
At the centre of this push is Boise, Idaho, where Micron is headquartered. There, it is pouring $50 billion into more than doubling the size of its 450-acre campus. Two massive fabrication plants are under construction. The first, known as ID1, is expected to produce its initial silicon wafers in mid-2027, manufacturing DRAM for high-bandwidth memory (HBM) chips, which are essential for advanced AI computing. Both fabs are slated to be operational by the end of 2028.
Each facility will span 600,000 square feet, among the largest clean rooms ever built in the United States. Preparing the ground alone required blasting through over 7 million pounds of dynamite. Construction crews have erected what resembles a temporary city to keep work running around the clock.
The scale is staggering, with 70,000 tonnes of steel per fab, almost rivalling the amount used in the Golden Gate Bridge, and 300,000 cubic yards of concrete, enough for four Empire State Building structures.
The AI boom ignites a manufacturing arms race
The frenzy is fueled by artificial intelligence. As companies such as OpenAI, Oracle, xAI and Anthropic unveil ambitious data centre plans, memory demand has outpaced supply. Advanced processors from Nvidia, Google, Broadcom, and Advanced Micro Devices require ever-faster, higher-capacity memory for both model training and inference.
The result has been a gold rush. Since April last year, Micron’s share price has risen more than sixfold, valuing the company at nearly half a trillion dollars. Gross margins have surged from 18.5% in early 2024 to 56% in its latest quarter, with expectations of 68% ahead, levels once unimaginable for a business long treated as a commodity supplier.
From commodity to critical infrastructure
Memory chips were historically vulnerable to brutal boom-and-bust cycles. After pandemic-era demand for PCs and smartphones collapsed in 2022, inventories piled up and valuations tumbled. Manufacturers slashed output to steady prices.
Then AI changed everything. Taiwan’s Commercial Times reported DRAM contract prices jumping more than 170% in a year. Circular Technology, a Massachusetts-based reseller, indicates DDR5 chip prices have climbed nearly 500% since September.
Micron can currently meet only half to two-thirds of demand from some major customers. Buyers are now seeking multiyear contracts to secure supply and avoid price shocks. Beyond Idaho, Micron recently broke ground on a $100 billion fab complex near Syracuse, the largest private investment in New York’s history. It also announced a $9.6 billion expansion in Hiroshima. Meanwhile, competitor SK Hynix is building a $13 billion fab in South Korea and a $4 billion complex in Indiana.
Competitive pressure in a high-stakes market
Despite soaring demand, competition remains intense. In February, reports suggested Micron’s HBM4 chips had failed to secure a supply role for Nvidia’s upcoming Vera Rubin AI servers. The company rejected those claims, stating it is already shipping HBM4 and has sold out its supply of both HBM4 and HBM3e through year-end.
For an industry once defined by volatility, memory has become central to the AI economy. Micron’s $200 billion wager signals that the race to supply the world’s data is only just beginning.
“I’ve been here for 28 years, and I’ve never seen anything so disruptive as AI,” said Scott Gatzemeier, the Micron vice president who is heading the company’s $200 billion U.S. expansion. “As we started to transfer from training to inference, the amount of data required just exploded, and we just didn’t have enough clean-room capacity to satisfy demand. We realised we had a huge problem.”
“Our business is on an extraordinary trajectory,” said Mark Murphy, Micron’s chief financial officer, at an investor conference on Wednesday. “On the supply side, we are doing everything we can to add capacity,” he said. “But there is no easy or fast way to get that done.”
“We’re nowhere near the end of the shortage,” said Brad Gastwirth, Circular’s head of global research. “I think it lasts through the end of 2026 and at least the first half of 2027.”
“Memory has gone from being a system component to being a strategic asset,” said Sumit Sadana, the company’s chief business officer. “The promise of AI is all ahead of us.”
