Samsung Electronics downplays concerns about neocloud firms and AI bubble
Shares of neocloud firms including CoreWeave, which rents out computing capacity to tech firms, have been hit by concerns that their debt-fuelled data center expansions could backfire if AI demand slows
Samsung Electronics, the world's leading memory chip manufacturer, dismissed worries about neocloud companies potentially contributing to an AI bubble, stating that they make up a modest 8% of the market, according to Kim Yong-kwan, president of Samsung's component business division. Speaking at a Morgan Stanley conference in Singapore, Kim assured that the company does not anticipate these concerns to significantly impact the broader AI growth.
Neocloud firms like CoreWeave, which provide computing resources to tech firms, have faced shareholder concerns due to their debt-driven data center expansions that may falter if AI demand slows. Furthermore, the recent resignation of Nvidia CEO Jensen Huang could further restrict funding for smaller, highly leveraged neocloud entities that have aided the sector's explosive growth, according to analysts and investors.
Samsung's Head of Foundry Sales & Marketing, Kang Suk-chae, described the present AI investment surge as structural rather than cyclical, citing long-term contracts between chipmakers and customers as providing better visibility into demand and supply, distinguishing it from previous speculative booms that led to overcapacity. He emphasized that Samsung's chip shortage situation is expected to endure in the foreseeable future, and stressed that the company's foundry business has diversified manufacturing facilities to mitigate supply chain risks and offers a comprehensive solution encompassing memory and advanced packaging.
Kang reiterated Samsung's plans to distribute shareholder returns through a blend of cash dividends and a "substantial" share repurchase and cancellation program.
Written by urgent.news from The Hindu - Sci-Tech's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.