Micron calls the memory shortage driven by artificial intelligence "unprecedented"

NVIDIA's supplier Micron Technology Inc. stated that the ongoing shortage of memory chips has further intensified over the past quarter and reiterated that this supply shortage will continue into next year due to the surge in demand for high-end semiconductors required for artificial intelligence infrastructure.

Micron's Executive Vice President of Operations, Manish Bhatia, stated in an interview, "The shortage we are currently facing is indeed unprecedented." This statement was made shortly after Micron held a groundbreaking ceremony for a $10 billion production facility in Syracuse, New York, and it also reaffirms similar predictions made by the company last December.

Bhatia noted, "The high-bandwidth memory required for AI accelerators has consumed a significant amount of capacity across the industry, leading to severe shortages of traditional memory products aimed at consumer devices such as smartphones and personal computers."

He also added that PC and smartphone manufacturers have begun to queue up to secure memory chip capacity beyond 2026, while demand from the autonomous vehicle and humanoid robot sectors will further drive up the demand for these storage components.

Market research firm Counterpoint Research predicted last December that global smartphone shipments could decline by 2.1% in 2025 due to the increased production costs and compressed capacity caused by the memory chip shortage. PC manufacturers like Dell Technologies have also issued warnings that their businesses are likely to be affected by the current memory shortage.

Benefiting from the AI boom, the stock prices of the three major players in the global memory chip industry—Micron Technology, SK Hynix, and Samsung Electronics—are expected to rise significantly by 2025. SK Hynix stated that its chip capacity for 2026 has already been sold out; Micron also revealed that its AI-specific memory chip capacity for 2025 has been fully booked.

To prioritize supply for strategic enterprise clients like NVIDIA, Micron announced last December that it would terminate its popular consumer-grade memory business under the Crucial brand. The enormous demand for memory chips in the AI industry has further accelerated Micron's capacity expansion efforts in both the U.S. and Asia.

The company stated that it will officially achieve large-scale production of DRAM wafers in the second half of 2027.

DRAM chips provide the computing environment for high-performance processors produced by companies like NVIDIA and Intel, and they are also a core component of the high-bandwidth memory required for AI accelerators to operate efficiently.

Bhatia stated in an interview on Friday, "The core mission of our production base in Asia is to continuously upgrade to the next generation of technology."

He also added that, in contrast, nearly all of the new wafer capacity will be established in the U.S. mainland Micron's $10 billion project near Syracuse, New York, plans to build four DRAM wafer fabs, each the size of 10 football fields. The project is expected to achieve its first wafer rollout by 2030.

The American chip manufacturer will also add capacity for two new fabs in Boise, building on its existing R&D facilities. The first new plant in Idaho is scheduled to start production in 2027, and the second plant is already in the planning stage. Additionally, Micron is upgrading and expanding its existing production base in Virginia.

All of the above expansion plans are an important part of Micron's commitment to "shift 40% of DRAM capacity to the U.S. mainland." In 2024, Micron successfully secured $6.2 billion in subsidies from the U.S. CHIPS Act, and during the construction of the plants, it will also benefit from a 35% tax credit. These policy supports lay a solid foundation for achieving its capacity relocation goals