Yorkville CEO Predicts AI Memory Shortage Could Last Until 2029, Impacting Micron
The CEO of Yorkville America Equities has warned that the shortage of high-bandwidth memory (HBM) for artificial intelligence applications could persist until 2028 or 2029. Steve Neamtz cited SemiAnalysis research indicating HBM accounts for over half the cost of AI accelerators, with major suppliers already committed to their 2026 output. New production facilities may not reach full capacity soon enough to meet demand. Micron (NASDAQ:MU), a key player in memory manufacturing and the largest holding in Yorkville’s new NRAM ETF, faces prolonged pressure as HBM is three times more resource-intensive than conventional memory. Neamtz noted that AI infrastructure bottlenecks are already affecting hardware design decisions at companies like Nvidia. The CEO also highlighted rising prices for mainstream DDR5 kits and surging memory stock valuations due to the shortage. Yorkville’s NRAM ETF includes not only DRAM/HBM producers but also storage firms, reflecting broader supply constraints across AI-related components.
Topics: AI HBM memory shortage Micron Yorkville ETF