As the axis of competition for artificial intelligence (AI) development shifts from large-scale learning to inference, the shortage of memory semiconductors essential for ‘inference’ is expected to continue for a considerable period of time. With the advent of an unprecedented ‘memory golden era’, the prices of memory semiconductors such as DRAM and NAND flash (NAND) are rising vertically. The market predicts that the memory shortage will continue until 2030 or later. According to DRAM Exchange, a semiconductor market research company, on the 12th, the average price of general-purpose NAND products for memory cards and USB was $28.8 at the end of last month, which is five times higher than the price at the end of last year ($5.7). The average price of general-purpose DRAM products for PCs also jumped to $21 last month, 2.3 times higher than the price at the end of last year ($9.3). Due to a shortage of memory supply, prices are rising 2 to 5 times in just half a year. In particular, as demand for NAND increased rapidly, a reversal occurred where NAND became more expensive than DRAM within half a year. DRAM stores data temporarily while AI is currently working.