The most powerful market bet of our time, artificial intelligence, is facing a harsh reality: the cost of money. Surging government bond yields to decade-high levels have triggered a global wave of selling in semiconductor stocks, as investors re-evaluate the cost of the massive infrastructure required for the AI boom.
Plunge in Asian and US Markets
The correction that began on Wall Street moved with even greater intensity into Asian markets. In South Korea, giants Samsung Electronics and SK Hynix recorded losses between 7% and 8%, dragging the Kospi index down by more than 5%. In Japan, the Nikkei fell by 2.6%, while the broader MSCI Asia Pacific index moved 2% lower.
In the US, the Philadelphia Semiconductor Index saw its worst session since late July, dropping 5%. Specifically:
- Micron Technology fell by 7%.
- Nvidia recorded a 2.3% decline.
- Sandisk and Western Digital lost 9% and 7.4% respectively.
Pressure from Bonds and Oil
The root cause of the turmoil lies in the fixed-income market. The US 30-year bond yield touched 5.34%, its highest level since 2007, dramatically increasing borrowing costs for so-called "hyperscalers." When the discount rate rises, the present value of future earnings—on which tech stocks rely—inevitably decreases.
"Higher interest rates and geopolitical risk are making investors less willing to pay a premium for this growth," said Yoon Yun, CEO of Fibonacci Asset Management Global.
Meanwhile, geopolitical uncertainty in the Strait of Hormuz is keeping Brent crude prices above $91. Expensive oil fuels inflation, limiting central banks' ability to cut interest rates and adding an extra layer of risk to the markets.
Warning Signs from the Debt Market
Jitteriness has also spread to the corporate debt market. On Tuesday, at least seven companies withdrew planned bond issuances due to adverse conditions. If financing remains expensive, ambitious programs for new data centers and AI energy infrastructure will face stricter financial scrutiny, threatening the pace of digital transformation.
