SK Hynix falls more than 10% as rate fears deepen the AI financing selloff
The retreat spread across chipmakers and equipment suppliers. One analyst argues that rising borrowing costs are not evidence of weaker AI demand.
Loading page…
The retreat spread across chipmakers and equipment suppliers. One analyst argues that rising borrowing costs are not evidence of weaker AI demand.
Listen to this story
Investors are treating the affordability of AI infrastructure as a growing risk: hotter U.S. inflation revived rate-hike concerns, while financing and construction costs weigh on expectations for future spending. The October 15 selloff spread across Asian chipmakers, equipment suppliers and broader tech stocks, but the moves do not establish that demand for computing is weakening. For AI investors, the key test is whether funding pressure eventually leads cloud providers to reduce spending, cancel contracts or pay less to rent GPUs.
South Korea’s Kospi fell as much as 6.5%, and Japan’s Nikkei 225 lost more than 3%; the reported moves were intraday, not confirmed closing results.
Bloomberg reported that U.S. semiconductor stocks lost $1.6 trillion in market value over six consecutive losing sessions.
Broadcom CEO Hock Tan and Intel’s chief executive warned that financing could constrain AI expansion alongside chip and electricity availability.
Asian technology stocks sold off on October 15, 2026, with SK Hynix falling more than 10%, Samsung Electronics losing 7.1% and SoftBank Group dropping 7.9%. CNBC linked the retreat to fears of further Federal Reserve rate increases and a more expensive AI infrastructure buildout. The pressure centered on financing, not a confirmed collapse in computing demand.
The Asian retreat followed Wednesday’s losses on Wall Street, where hotter-than-expected U.S. inflation data fueled concern about additional rate increases. The Nasdaq Composite fell 1.88% to 26,692.87. The S&P 500 lost 1.11%, while the Dow Jones Industrial Average slipped just 0.09%, leaving technology shares at the center of the decline.
Thursday’s selling reached several layers of the hardware industry. In Japan, semiconductor-testing equipment maker Advantest fell 6.9%, and chipmaking-equipment supplier Tokyo Electron lost 5.4%. Renesas Electronics also declined 5.4%. South Korean losses included LG Innotek, down 8.1%, Seoul Semiconductor, down 7.7%, and Samsung SDI, down 6.8%.
The damage extended to national benchmarks. South Korea’s Kospi fell as much as 6.5%, while Japan’s Nikkei 225 dropped more than 3%. Bloomberg’s report, published by CNBC-TV18, put the accumulated U.S. semiconductor losses at $1.6 trillion in market value over six consecutive losing sessions. These Asian figures describe trading-session moves, rather than confirmed closing results.
The concern is the cost of paying for AI infrastructure. Higher borrowing costs could weigh on the large capital investments needed to build it, pressuring companies that benefited from the spending boom. Bloomberg also identified stretched valuations as part of investors’ unease, alongside interest rates and construction costs.
Chip executives had already raised that risk. On Monday, Broadcom CEO Hock Tan warned that financing could become a bigger obstacle to AI expansion than shortages of chips or electricity. Intel’s chief executive also warned that financing could become a major headwind. Those comments put access to funding alongside the physical resources needed for expansion.
CoreWeave added to the anxiety: its shares plunged as its bonds fell to levels Bloomberg described as typically indicating distress. That description is a market warning, not a declaration that the company has defaulted. Life Asset Management CEO Kang DaeKwun saw depressed South Korean valuations as a possible entry point, but cautioned that technology-funding problems needed to be resolved.
Higher yields raise the cost of financing these builds, and investors are right to focus on balance sheets and contract structure, but rates are a financing question, not a demand question.
Dan Ives, technology analyst at Yorkville Ives, in a note quoted by CNBC
Ives offered the clearest competing interpretation. He described demand for AI computing as robust and called the pullback a buying opportunity, arguing that the infrastructure boom remained in its early stages. His position was not that financing costs were irrelevant. It was that a change in demand would carry more weight than a change in rates.
He identified three developments that could change that view. These were warning signs to watch, not claims that they had already occurred:
Company results have not yet settled investors’ concerns. Strong earnings from Samsung and ASML failed to calm the market, Bloomberg reported. Taiwan Semiconductor Manufacturing Co.’s earnings, due later on October 15, were the next company-specific checkpoint it identified. The distinction remains important: healthy earnings can coexist with investor doubts about the affordability and durability of future infrastructure spending.
Loading discussion...
Join the conversation
Explain which risk would matter more to your confidence in the buildout.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.