Nvidia CEO dismisses 2029 memory oversupply worries, eyes continued LPDDR use
Nvidia chief Jensen Huang told investors that high-bandwidth memory will stay in demand and the company will keep deploying large volumes of LPDDR chips in its data-center products.
At the GTC Berlin event, Nvidia CEO Jensen Huang reassured the market that high-bandwidth memory (HBM) will remain essential for years and that the firm will continue to rely heavily on LPDDR memory in its data-center offerings. His remarks followed a recent South Korean column that warned of a possible memory surplus by 2029, which had contributed to a decline in AI-sector equities. Elon Musk responded to the column by predicting massive RAM production in China for 2028.
Huang also addressed speculation about a partnership with ChangXin Memory Technologies, saying Nvidia is open-minded and works with Chinese suppliers across the supply chain. He credited U.S. regulations for accelerating China’s semiconductor initiatives and affirmed that American tech firms are prepared to compete with any challenger. The comments arrived as AI stocks have been falling for two weeks, marking the longest consecutive drop in the semiconductor index since 2008.
