Micron Technology stock plunged 19.5% over five days following Google's TurboQuant announcement, while SK Hynix and Samsung also saw significant declines. Google TurboQuant, a new AI model compression algorithm, claims to slash memory usage by 6x and speed up inference by 8x. This breakthrough combines PolarQuant and Quantized Johnson-Lindenstrauss (QJL) techniques to drastically reduce key-value cache size to just 3 bits.

The algorithm's ability to accelerate GPU runtimes without accuracy loss positions it as a major industry shift. Industry analysts note this stock reaction signals immediate market concern over potential disruption to DRAM demand.