I am quite interested in the topic, and I am sure they will be even more so. Now that we have discussed memory, perhaps we could talk more about that since I have spent more time on logic. The whole memory cycle and where we fit into that cycle, and how it might have changed or been extended. Could you explain if you think the AI-driven demand will make HBM, which we have seen explode, structurally more important and a more significant part of the industry going forward?

In the past, memory was treated very much like a commodity. It was more about operational excellence than necessarily innovation. The people who made money in the memory business were those who had a combination of scale and either the best ability or luck in matching supply with demand. Their capital expenditures were carefully matched to industry demand, and making the tech node inflection and changeovers correctly was crucial.

I am quite interested in the topic, and I am sure they will be even more so. Now that we have discussed memory, perhaps we could talk more about that since I have spent more time on logic. The whole memory cycle and where we fit into that cycle, and how it might have changed or been extended. Could you explain if you think the AI-driven demand will make HBM, which we have seen explode, structurally more important and a more significant part of the industry going forward?

Samsung was one of the masters in this, which drove them to be one of the top suppliers. They still are among the top suppliers in commodity items such as DRAM and SRAM. However, they slipped in the HBM introduction, where SK Hynix really excelled.