Asia’s technology stocks took another beating on Friday as a brutal sell-off in the AI trade swept across the region.

Some of the chipmakers and technology companies that powered this year’s stock-market rally were among the hardest hit.

South Korea’s benchmark Kospi index plunged 9% in early trading, triggering a trading halt. Japan’s Nikkei 225 sank 7%.

In Seoul, memory chip giants Samsung Electronics and SK Hynix fell sharply. Samsung dropped as much as 11%, while SK Hynix — a key supplier of high-bandwidth memory chips used in AI processors — lost nearly 14%.

Japanese technology stocks were also hammered. Memory chipmaker Kioxia slumped 15%. Chip-testing equipment maker Advantest fell about 16%, and technology investor SoftBank Group tanked 15%.

Shares of Taiwan Semiconductor Manufacturing Company, a key supplier to Nvidia and Apple, fell as much as 10%. Markets in Hong Kong and China were also lower in early trade.

The rout followed a bruising session on Wall Street on Thursday, when the Nasdaq 100 entered an official correction. The sell-off came after Democrats swept the US midterm elections.

Investors fear the Democratic victories could bring tighter regulations of data centers and AI development, threatening the spending boom that has fueled the stock market rally and helped drive US economic growth.

Asia is particularly exposed to the reversal after a year of blockbuster stock market gains fueled by enthusiasm for AI.

South Korea, Taiwan, and Japan are home to some of the world’s biggest suppliers of the chips, memory, and equipment needed to power the technology.

“The sharp correction is partly a reassessment of future US infrastructure demand, but also reflects how heavily investors have positioned in the same beneficiaries of the AI boom,” wrote Charu Chanana, Saxo’s chief investment strategist, in a note on Friday.

The sell-off also comes amid concerns about the sustainability of AI-infrastructure spending, compounded by investors’ crowded positions in technology stocks.

“The market is repricing future expectations rather than responding to a sudden collapse in current demand,” Chanana wrote.

Investors are increasingly looking for evidence that massive AI spending will pay off, rather than simply rewarding companies for expanding capacity. That could favor businesses with strong finances and established customers over those relying on outside funding to fuel growth, she added.

“This is not necessarily the end of the AI trade, but it may be the end of rewarding every company associated with AI equally,” Chanana wrote.