Confident to say that concerns about the AI rally have been erased after Micron earnings and SK Hynix's US listing plans?

For now, the Micron earnings are very consistent with our thesis—as much as it's been a fierce rally, it is backed by earnings and fundamentals, and valuations are still reasonable.

To what degree can we say 'buy the dips' has been reinforced for AI?

At some point that game will be over, but the transfer of wealth from hyperscalers' massive orders and the earnings visibility for semiconductor players is something we haven't seen even during the TMT crisis.

Can you see a scenario where hyperscalers build their own memory chips and cut out the middleman?

There will always be second, third, fourth players—Chinese catch-up happens quicker than expected. But the evolution of semiconductors means leaders become stronger each cycle. I foresee SK Hynix, Samsung, and Micron taking the high-margin cutting-edge segment while others take commodity (mobile phones, PCs)—but that takes 2-3 years. Concerns about hyperscaler funding sustainability are still 1-2 years away. The visibility semiconductor producers have on their earnings is very compelling.

Is SK Hynix's US listing a defensive play against US trade protectionism?

That might be a couple of steps further away. Everyone in AI space is trying to raise money—maybe signaling concern about the funding environment a year or two away. On protectionism, I do worry a bit—three companies cornering the entire AI chip market. But putting barriers up on a scarce commodity whose prices are rising so fast seems more about securing inventory than making it a nationalistic issue.

When retail investors chase stocks on FOMO, that's usually a signal we're near the top. Is that a fair assessment?

That's exactly what happened during the TMT bubble. However, this is a rare case of retail investors chasing a handful of stocks with very strong balance sheets, backed by earnings, with order visibility of 1-2 years out. As much as price action looks like a bubble, we are probably only halfway through this cycle. That gives plenty of leeway for expectations to eventually surpass fundamentals.

Is labor friction a threat we might see more of at SK Hynix and Samsung?

For this year, the discussion is over, but they're already trying to establish profit-share schemes for next year. The concern is other Korean companies might follow their leadership on collective bargaining, putting a burden on labor flexibility. Countries with adaptable labor will win in AI transition—Korea seems to be going the other way. But earnings are so powerful that this is a secondary consideration for now. When the cycle declines, we'll talk about it more seriously.

Can you give year-end targets for KOSPI, SK Hynix, and Samsung?

All three are essentially the same question since those two stocks make up over 50% of KOSPI. I feel comfortable with KOSPI at 10,000, which would probably mean both Samsung and SK Hynix have a comfortable 30% upside from here.