AI Infrastructure Build-Out Benefits Quality Names

SPEAKER_1 (0:00)

To discuss three standout industries in the current environment, let's welcome in our next guest. Joining us, Jed Ellerbroek, Portfolio Manager at Argent Capital Management. Jed, great to have you on. Where are you seeing opportunity right now in the market?

Jed Ellerbroek (0:14)

Hey, Marlee, great to be with you today.

I think one area where there's opportunity are the cloud computing giants. Microsoft, Amazon and Google, I think, are all pretty compelling stocks. Those valuations have drifted down over the last year. Investors, you know, it seems like we rotate between thinking these companies are AI winners one month, and then you look six months later, investor sentiment has turned quite a bit. But all the while, these companies are demonstrating, accelerating revenue growth. Their valuations are becoming more appealing. As I said, in the margin compression that investors feared related to AI a couple of years ago has not come to pass. Margins are actually going up at Amazon and at Google.

SPEAKER_1 (0:53)

And so as we look at these names, I mean, they're developing their own chips. How important is Custom Silicon to the long-term economics of AI for Microsoft, Amazon, and Google?

Jed Ellerbroek (1:07)

Exceptionally important. These companies are handling lots and lots of workloads within their data centers for customers. Some of those workloads are very, very demanding and they require the top performing Nvidia chips. Others are more ruder standard and perform very well being used through their custom chips which are a little bit more specialized and a little bit less general purpose oriented and capable.

A mix of both is very, very important for all three companies.

The other benefit of the custom chips is they're designed in-house, and so they carry a higher margin profile. That's one of the key reasons why you see Amazon and Google's margins performing so well. Microsoft's chip program is earlier in its deployment and its maturation process, and we will expect margin improvement to come from that over the next couple of years as they get it more widely deployed.

SPEAKER_1 (2:02)

Jed, one of the interesting points in your thesis is that these companies benefit whether the winning models are open source or closed source. There's been a lot of discussion around this, especially with Nvidia's acquisition of Hugging Face. Explain that to me.

Jed Ellerbroek (2:19)

Yeah, the open AI and Anthropic, obviously two private companies that have the leading closed source models. Investors are very excited about those companies. One of the risk factors is that open source models, which are far cheaper and easier to customize, kind of encroach on or catch up to the capabilities of closed models. We've seen a little bit of that over the last six months. The most capable open AI and Anthropic models in the market, the open source models have been kind of closing ground, you would say.

But the cloud computing giants, Google, Amazon, Microsoft, they generate revenues via usage and consumption on their cloud computing platforms, and they charge less per task to use it. An open source model versus closed, but there are lots of additional services like security and database, et cetera, that go along with the use of open source models. The profitability of the two is really pretty similar for these big datacenter operators. So they're very happy to shift customers to the cheaper, more customized open source products.

SPEAKER_1 (3:26)

All right, another area you like is aerospace. And I want to walk through this one because it's struggled recently. I mean, I'm just looking at the ITA ETF right now, down more than 10% over the last month. In year to date, it's only up 3%.

So how are you looking at aerospace here? And what's the market getting wrong?

Jed Ellerbroek (3:45)

Yeah, the market is very focused on number one, the war in Iraq and the Iraq, I meant to say Iran, sorry.

The impact of oil prices from the war with Iran, and then also the increase in Boeing's production rate. Both of those are risk factors for aerospace aftermarket businesses like Transdime, GE, Heiko and others. We think those stocks are increasingly attractive. Their valuations have become more appealing as this year is going on and those stocks have underperformed. They're not underperforming because earnings estimates are going down. They're not underperforming because their financial results are weak, or because their revenues are declining. The opposite is true in all of those cases. Investors are just fearful that higher oil prices are going to lead airlines to cut their flight schedules, which means less demand for aftermarket services, the likes these companies provide. Then the other one, Boeing increasing their production rate, more new airplanes means less wear and tear on the old ones, perhaps. I think what investors are missing, though, is that all of the airplanes are being used today. The new airplanes that Boeing is delivering, that higher production rate, is going to go in service of new routes and new flight tracks for the various airline companies around the world. I think that demand for aftermarket services is durable. I think these companies are going to keep growing, growing at an above GDP rate. They also have pricing power, because the aerospace industry has significant regulations on a part by part basis, and it's pretty difficult for airlines to toggle from one to another.