AI Funding Concerns Hit Asia’s Hardware Suppliers
A regional AI hardware basket fell 11% last week as investors questioned whether infrastructure spending can keep pace with borrowing costs.
Key takeaways
- A Goldman Sachs basket of regional AI hardware stocks fell 11% last week.
- Nan Ya Plastics fell 12% from a record high, while Samsung Electro-Mechanics dropped 23%.
- CoreWeave shares lost 25% over the week as investors questioned AI-related borrowing.
- The article reported no clear evidence that financing strains had weakened orders.
A Goldman Sachs basket of Asian AI hardware stocks fell 11% last week, its steepest decline since mid-July, as investors questioned whether AI infrastructure spending can withstand higher financing costs. The drop came despite an upbeat revenue outlook from Taiwan Semiconductor Manufacturing Co.
The pullback reached some of the region’s biggest recent winners. Nan Ya Plastics fell 12% from a record high, Unimicron Technology lost 14% in a week, and Samsung Electro-Mechanics dropped 23%.
Financing concerns spread
Investors are looking beyond demand for chips to whether hyperscalers and neo-cloud operators can fund the data centres and other infrastructure needed to deploy them. Oil’s rise to US$110 a barrel, its first time at that level since May, added to inflation concerns and worries that borrowing costs could stay elevated.
The pressure was also visible in the United States. An eight-session decline in semiconductor shares, the sector’s longest since 2008, erased about 16% from its benchmark, while the Nasdaq 100 had its worst week since April 2025.
CoreWeave’s bonds fell to levels signaling distress last week as investors questioned whether AI demand and revenue would support its borrowing. Its shares dropped 12% on Friday, capping a 25% weekly decline, according to the article.
Orders remain a key test
The risks may be greater for suppliers further down the hardware chain. Kevin Net, a portfolio manager at Financiere de l’Echiquier, said project delays could have a bigger impact on downstream companies than on the core computing ecosystem.
Gartner has warned that more than US$3.5 trillion of AI infrastructure spending could be at risk of impairment by 2030 if investment outpaces the revenue needed to justify it. Broadcom CEO Hock Tan has also said capital, rather than shortages of memory chips or manufacturing capacity, could become the main constraint on AI growth.
Still, the article said there was little evidence that financing strains had translated into weaker orders. Kunjal Gala of Federated Hermes said demand for AI remained strong and that he was not worried about the chip sell-off.
This article was produced with the help of AI technology. Based on reporting by the outlets listed above.
