Are Index Funds Making the Market “Irrational”?

**Jon Quast** (0:02)
Are index funds making the market irrational? You're listening to Motley Fool Hidden Gems Investing.
Welcome to Motley Fool Hidden Gems Investing. I'm Jon Quast, and I'm joined today by Foolish Contributors, Matt Frankel and Rachel Warren. Before you dive into today's episode, do yourself a favor. Head over to news.fool.com and sign up for Breakfast News. It's a free daily email from the Motley Fool, landing in your inbox by 7.30am every morning, packed with the stories long-term investors actually want to know about.
We're going to tackle some topics today on today's show. We have actually two topics about ETFs from our mailbag. Something like that you wouldn't necessarily find in Breakfast News, but you might find something in Breakfast News similar to the story we're leading with here, and that is the topic that we have regarding SK Hynex. Now, SK Hynex is a computer memory company based in South Korea, and it's already publicly traded there in South Korea, but as early as this week, it does plan to list some American depository shares here in the US, or ADR. It's going to trade under the ticker symbol SKHY, and it's targeting to sell nearly 178 million ADR shares. That will hopefully raise roughly $28 billion.
It needs the money to build new factories and furnish them with chipmaking equipment. Now one could say with AI, GPUs aren't holding anything back. It's the computer memory finding that specifically that high bandwidth memory. That is what these companies need right now. That is what we are not making enough of. And my first question here to you, Rachel, is what exactly is high bandwidth memory and who makes it?

**Rachel Warren** (1:52)
So, high bandwidth memory or HBM as it's commonly known for short, that's basically the ultimate data superhighway for artificial intelligence. So, traditional memory chips sit far away from the computer processor that is creating major data traffic jams, particularly in the age of intensive AI applications. So, high bandwidth memory aims to solve this by stacking memory chips vertically, like a skyscraper if you will, and placing that entire stack right next to the main processor. So, this lets massive amounts of data travel, in some cases up to 10 times or more faster while using way less power. And so, in short, we're at a time where AI processors are constantly starved for data, high bandwidth memory is the only memory fast enough to keep them fed. Now, right now making these chips, it's so complex, it's so expensive. There's only three companies in the world that control this entire market. ST Hynex is the undisputed king of the mountain of this space. They command over 50% of the market as the primary memory supplier for NVIDIA's AI chips. Now, the other two players are Samsung, global memory giant also based in South Korea, and Micron, the only major player based in the US. So, because building these high tech factories requires billions and billions of dollars, ST Hynex is coming to Wall Street to secure the cash they need to stay ahead in this very, very intense race.

**Jon Quast** (3:14)
I think that so many investors missed the whole memory trade because historically memory is such a commoditized market and there's so many fears related to historic patterns when it comes to this commoditization.
But right now, it is enjoying these companies, specifically SK Hynex, enjoying these incredible business economics.
I would say it's very smart for it to capitalize on the trend right now. Go public here in the US., raise that capital that it needs. So it's good IPO timing, no doubt. I am curious here about long-term shareholders. Do either of you think that SK Hynex is a good investment when it comes public, or are you looking at that $28 billion that it's going to be raising and saying, maybe there's a secondary beneficiary here because that money is going to go somewhere?

**Rachel Warren** (4:05)
Yeah. There's no denying that SK Hynex, they're riding an incredible wave and timing this US listing during what is essentially peak AI euphoria. I think it's a brilliant move from a corporate perspective. Now, I am not planning to buy shares at any point in the near future, and my hesitation comes down to a few things. Obviously, there's capital intensity, but there's also the long-term cyclical risk. Building and equipping these facilities requires an astronomical amount of cash as I was discussing. As a shareholder, my concern is that buying in today means you're betting that AI demand will remain hot enough or long enough to absorb all this new capacity. Even if it does, there is a lot of excitement that's baked into the stock right now.