NVIDIA (NVDA.US) supplier Micron Technology (MU.US) stated that the persistent memory chip shortage has intensified over the past quarter and reaffirmed that this supply tightness will continue beyond this year due to surging demand for high-end semiconductors required for AI infrastructure. "The shortage we are seeing now is truly unprecedented," said Manish Bhatia, Micron's Executive Vice President of Operations, in an interview. This statement came shortly after the chipmaker held a groundbreaking ceremony for its $100 billion production base in the suburbs of Syracuse, New York, further reinforcing a similar prediction the company made in December.
Bhatia pointed out that high-bandwidth memory (HBM), essential for manufacturing AI accelerators, is "consuming a vast amount of the industry's available capacity, leading to significant supply shortages for traditional sectors like mobile phones and personal computers (PCs)." He added that PC and smartphone manufacturers have already joined queues, attempting to lock in memory chip supplies for 2026 and beyond, while autonomous vehicles and humanoid robots will further drive up demand for these components.
Last Friday, Chinese media reported that major smartphone manufacturers, including Xiaomi Group, Oppo, and Shenzhen Transsion Holdings, are cutting their 2026 shipment targets due to rising memory costs. Industry tracker Counterpoint Research estimated in December that global smartphone shipments could decline by 2.1% this year, as memory chip shortages drive up costs and squeeze production.
PC manufacturers, including Dell Technologies, have also warned that they could be affected by the ongoing shortages. Benefiting from the AI boom, the stock prices of the world's three largest memory chipmakers—Micron, SK Hynix, and Samsung Electronics—have surged significantly in 2025.
SK Hynix stated that its entire chip production capacity for 2026 is already sold out, while Micron also indicated that its AI memory semiconductors for this year have been fully booked. The AI boom has significantly boosted the performance of the three major players in the memory industry.
To prioritize supply for strategic enterprise customers, including NVIDIA, Micron announced in December that it would discontinue its popular Crucial brand consumer memory business. The AI industry's nearly "insatiable" demand for memory chips has also increased the urgency for Micron to expand its manufacturing scale in the US and Asia.
Last Saturday, Micron announced plans to spend $1.8 billion to purchase a site with existing factory buildings in Taiwan, China, a key production hub for the chipmaker. This move significantly shortens the timeline for Micron to bring new factories online. The company stated that it will begin producing DRAM wafers in meaningful volumes in the second half of 2027.
DRAM provides the operating environment for complex processors from NVIDIA and Intel to perform computations and is also core to the high-bandwidth memory required for AI accelerators to function optimally. "What we are doing at our Asia base is to continue the transition to next-generation technology," Bhatia said in the interview. He added that, on the other hand, the additional wafer capacity will be almost entirely realized within the United States.
Micron's $100 billion project near Syracuse plans to build four DRAM wafer fabs, with each facility roughly the size of ten football fields. The first wafers are expected to come off the line around 2030. The US chipmaker is also adding capacity with two wafer fabs next to its existing R&D facility in Boise. The first fab in Idaho is planned to start production in 2027, with a second factory in the planning stages.
Furthermore, the company is modernizing and expanding its existing manufacturing facility in Virginia. These plans are part of the company's commitment to shift 40% of its DRAM manufacturing to the US mainland. The achievement of this goal is supported by the $6.2 billion in CHIPS Act funding the company received in 2024, along with an investment tax credit, now increased to 35%, that it can utilize during construction.
Comments