Memory manufacturers are increasingly shifting production towards AI-oriented products, directing a growing share of global memory capacity to data center applications. This trend has led to concerns among U.S. trade associations, representing telecommunications, automotive, and medical device industries, which urged the Trump administration in June 2026 to address emerging memory chip shortages.
Global memory sales increased nearly 79 percent in 2024 to $165 billion, projected to exceed $223 billion in 2025, driven by demand for high-bandwidth memory (HBM) from AI infrastructure. Major technology firms like Meta, Microsoft, Amazon, and Alphabet are projected to increase annual spending on AI-capable data centers from $217 billion to roughly $650 billion between 2024 and 2026. OpenAI's Stargate project alone reportedly secured agreements for up to 900,000 DRAM wafers per month in 2025, equivalent to about 40 percent of global DRAM output. Samsung, SK Hynix, and Micron, which account for over 90 percent of global DRAM production, have expanded HBM output, with data centers projected to consume roughly 70 percent of worldwide memory output in 2026. These shortages are expected to persist through at least 2027 due to the multi-year investment required for new fabrication facilities. Access to memory components is becoming a critical determinant of industrial competitiveness and AI deployment, necessitating continued investment in manufacturing and advanced packaging capacity, and strengthening supply chain resilience.
No comments:
Post a Comment