Micron Technology poured the first concrete at its sprawling New York megafab on July 9 and used the moment to make a far bigger declaration: the memory maker will now spend more than $250 billion in the United States through 2035, up from a prior $200 billion commitment, to keep pace with an AI boom that is straining the world's supply of memory chips.
The milestone in Clay, New York, arrived more than a full quarter ahead of schedule, less than six months after Micron broke ground in January 2026. It marks the transition from site preparation to vertical construction on what the company and state officials call the largest private investment in New York history.
"As America celebrates its 250th anniversary, data and memory are foundational to the modern economy — and Micron is increasing our U.S. investments to more than $250 billion through 2035 to meet that moment," said Sanjay Mehrotra, Micron's chairman, president and CEO, who hosted the concrete pour alongside Commerce Secretary Howard Lutnick and New York Governor Kathy Hochul. "Reaching this milestone ahead of schedule reflects the speed and determination behind this project."
Building America's memory footprint
The expanded spending is meant to push Micron toward a long-term goal of producing 40% of its DRAM inside the United States. The New York campus, planned for up to four fabs, is projected to generate 50,000 jobs statewide, including 9,000 direct Micron positions. To date the company has directed roughly $675 million to New York-based contractors and suppliers, with more than 80% of on-site workers being state residents. Bechtel is leading engineering and construction, with Jacobs on design and Gilbane on early site work.
New York is the cornerstone, but not the whole plan. Micron expects first wafer output from its first Idaho fab in mid-2027 and a second in late 2028, and earlier this year began initial production of 1-alpha DDR4 at its Manassas, Virginia site for automotive, industrial, medical and defense customers. Together the projects are expected to create more than 90,000 jobs. Micron also unveiled a separate plan to invest up to $3 billion in the domestic semiconductor supply-chain ecosystem.
"This milestone in Central New York shows Micron's U.S. manufacturing strategy moving from planning to meaningful local impact," said Manish Bhatia, Micron's executive vice president of global operations. "As we build the capacity, workforce and supplier base needed for the AI era, we are creating opportunities for New York businesses, skilled trades and communities to grow with us."
The buildout leans heavily on Washington. Senator Charles Schumer credited a $6.1 billion CHIPS Act grant and billions more in investment tax credits for making the project possible, while Lutnick framed the announcement as proof of the administration's manufacturing agenda, saying the commitment would create "nearly 100,000 jobs."
Locking in demand before the shortage bites
The capital surge is underwritten by a wave of long-term contracts. In the days before the concrete pour, Micron signed a five-year Strategic Customer Agreement with Ford, following a similar deal with General Motors on July 1. Those automotive pacts are two of 16 such agreements the company disclosed on its fiscal third-quarter earnings call, spanning four large hyperscalers — Meta among the cloud giants racing to secure supply — down to nine smaller automotive suppliers.
The contracts are unusually binding. Structured as five-year, take-or-pay deals running from 2026 through 2030, they carry fixed volumes and rigid pricing, and Micron says it has collected $18 billion in cash deposits plus $4 billion in letters of credit — $22 billion in guarantees that now backstop roughly 40% of its business. The urgency is visible in prices: DRAM has climbed about 70% since December as AI data centers vacuum up capacity, forcing automakers to bid directly against hyperscalers for the same scarce chips.
The HBM race that changes everything
Behind the numbers sits high-bandwidth memory, the stacked DRAM that feeds Nvidia's AI accelerators. Micron says its entire calendar 2026 HBM output, including its new HBM4, is already sold out. The company projects the HBM market will roughly triple, from about $35 billion in 2025 to near $100 billion by 2028, a compound growth rate of about 40%.
That prize is fiercely contested. SK Hynix remains the dominant supplier, holding roughly 56% to 62% of the HBM market depending on the quarter, having locked up much of Nvidia's Hopper and Blackwell demand. But Micron has quietly overtaken Samsung on some allocations to claim second place at around 21%, shipping HBM4 samples rated up to 11 Gbps. The New York and Idaho fabs are Micron's bid to convert that momentum into durable US-based capacity rather than a one-cycle gain.
What to watch next
The execution risk is real. Micron warned it will stay "disciplined" and responsive to the market, a reminder that memory is famously cyclical and that today's shortage can flip to glut. Watch for HBM4 qualification wins at Nvidia and whether Samsung claws back share; the pace of Idaho's 2027 ramp; how much of the $250 billion is contingent on continued CHIPS-era incentives; and whether the take-or-pay contracts hold if AI capital spending cools. For now, Micron is betting that demand outruns supply well past 2027 — and pouring concrete to prove it.
"As America celebrates its 250th anniversary, data and memory are foundational to the modern economy — and Micron is increasing our U.S. investments to more than $250 billion through 2035 to meet that moment."— Sanjay Mehrotra, Chairman, President and CEO, Micron Technology