On August 13, during an exclusive interview at SK Hynix's headquarters in South Korea, Chairman Chey Tae-won stated that memory demand is experiencing explosive growth, and the most severe 'memory shortage' is expected next year. He pointed out that the demand for memory within the AI ecosystem has surpassed the company's supply capacity, with all clients requesting nearly double the supply, while capacity expansion will take four to five years. Chey emphasized that Nvidia is indispensable in AI infrastructure, but SK Hynix is not reliant on a single client, also supplying to Google, Microsoft, and others. With the surge in demand for AI servers, HBM (High Bandwidth Memory) is in short supply, prompting tech giants to rush to sign long-term agreements. Chey predicted that the use of AI agents could increase 77-fold within five years, with memory demand continuing to grow significantly.