Alphabet’s Google has unveiled its KV cache quantization compression technology, TurboQuant, promising dramatic reductions in memory usage for AI inference. While the innovation has captured global attention, South Korea’s academic and industrial sectors remain skeptical about its practical feasibility, even as they firmly expect AI inference to continue driving substantial growth in memory demand.