The AI trade, which includes semiconductors and memories, is showing signs of fatigue as investors reassess whether the extraordinary spending boom in chips and data centers can be sustained.
Semiconductor and memory stocks like Micron Technology (MU) and Sandisk (SNDK) came under heavy pressure on Tuesday, after Samsung Electronics (005930) reported record second-quarter earnings but missed revenue estimates.
Still, shares fell nearly 7%, extending a broader sell-off among AI-linked chipmakers. Concerns are growing that hyperscalers could slow spending on AI infrastructure.
Meanwhile, rival SK Hynix is down 25% from its all-time high ahead of its US IPO this week, a deal that is also shifting investor capital away from existing chip stocks.
The weakness comes after a spectacular rally in AI infrastructure stocks this year, with Sandisk up more than 525%, Micron gaining more than 120% and SK Hynix up about 225% in 2026.
Adding to the changing narrative, China’s Zhipu AI, one of the country’s leading AI startups, is exploring a custom AI chip as demand for its open-source GLM models rises, highlighting the rise of lower-cost AI ecosystems built around domestic hardware rather than cutting-edge American chips.
The move comes just weeks after SpaceX’s successful initial public offering and amid elevated valuations for AI-related stocks. Investors are increasingly wondering whether the next phase of AI will increasingly require GPUs and high-bandwidth memory, or whether more efficient models will reduce demand for the infrastructure that has fueled the AI rally.
Over the past year, bitcoin and the broader crypto market have been hit by AI trading, and if investor enthusiasm for AI continues to fade, cryptocurrency bulls could see capital shift back into digital assets.
