Shares of major tech chip makers, including Micron Technology, Sandisk and Samsung dropped sharply this week following news of a new AI algorithm from Google. Micron fell 6.6% and Sandisk 9.3% on Thursday, continuing losses from Wednesday.

The decline was sparked by Google’s announcement of TurboQuant, an algorithm designed to compress AI model data, reducing memory requirements by up to six times while increasing processing speed by as much as eightfold without sacrificing accuracy. Investors worried this efficiency could slow long-term demand for memory chips used in AI systems.

Evercore analyst Amit Daryanani said TurboQuant “highlights a path toward materially reducing memory intensity in AI workloads,” which could pressure future demand for DRAM and NAND chips if widely adopted.

However, industry experts caution against overreaction. TurboQuant research has been publicly available since April 2025, and similar memory-efficiency efforts are underway across major AI and cloud companies. Jordan Klein, an analyst at Mizuho, noted that if the technology were already revolutionary at Google, the company would not have published it.

Experts also highlight a potential upside: improved efficiency could increase overall AI demand. More capable, efficient AI models may drive higher usage and investment, a pattern seen after China’s DeepSeek AI breakthrough in 2025. Sandisk CFO Luis Visoso said TurboQuant could boost returns on infrastructure spending, encouraging even more hardware investment.

Wells Fargo analyst Aaron Rankers echoed the sentiment, calling the improvements “another example of AI’s push for greater efficiency gains.” Analysts maintain that while TurboQuant briefly rattled markets, it is unlikely to derail the long-term growth of chip makers.