Episode 105 | Marvell’s Growing Role in the AI Infrastructure Boom

Rodney Lake (0:01)

Thank you for joining Market News with Rodney Lake. This is a regular program for the GW Investment Institute, where we talk about timely market topics. I'm Rodney Lake, the Director of the GW Investment Institute. Let's get started.

Welcome back to Market News with Rodney Lake. I'm your host, Rodney Lake. We're coming to you from the George Washington University, the George Washington University School of Business, Duques Hall, Duques Family Studio, right here in the heart of Foggy Bottom. This is a GW Investment Institute podcast. Remember, educational purposes only. Today, we're going to go over a company. We have not gone over this company before using the Business Management Price Valuation and Balance Sheet, BMPB, GW Investment Institute, Brainwork. All right. What you know and I hope love, put the comments down there if you'd like.

Thanks to everybody who's been watching, and certainly send us comments and ideas and whatever you have there. Love to hear from you and please check in. But thank you so much if you've been watching on YouTube, listening on Apple or Spotify. Thank you so much. All right. Let's get into it. The company is Marvell Technology. You may not have heard of this company. If you're watching the data center world, you're watching the AI world, you probably have heard of this company.

We have not talked about this company before, but given everything that it's happening, I thought it's time to check in on Marvell and see what's really going on. The ticker for Marvell, if you're not following it, is MRVL. That is the ticker. Marvell Technologies, what do they do? Well, this is a Fabless, so meaning they don't have a fabrication facility, designer of data infrastructure chips.

We're not going to get into all the specifics around that. We're going to broadly talk about each of these categories, but we want to introduce this. Maybe we'll do a deeper dive one day on the more specifics around here. But the important part and why I wanted to talk about this is that it makes chips for the data centers and communications. Really the part that's been growing is this AI story and these custom chips. That is really... Think of Amazon as a customer and their TPUs. I'm sorry, their Tranium chips. That's Google TPUs.

Their Tranium chips. This business has been taking off. And if you look at... And certainly, this is on the optics and the interconnection pieces. And so, if you look back and somebody thought that that was very important, was a person named Jensen Wong. We talk about in the video all the time. CEO of Nvidia. And they put a $2 billion investment in Marvell back in March earlier this year in 2026 We're recording this towards the end of September. So, I think it's time for a follow up here to talk about it. We haven't talked about it on the episode, so follow up for me to be thinking about this and see really what's going on. Now, again, this really sits at the center of the data center business, and it is part of what's happening and what is new on the optic side and moving data around. This is really a data infrastructure play, but this is a fabulous chip maker in that world. I mean, they don't make all their own chips, and we'll talk about that, and you're familiar with another company that we talk about Taiwan Semiconductor. So, we'll talk about that as we talk about the business. But, as we typically do, let's get some stats underway here for the business to describe it from a number standpoint, so we get a sense of, okay, well, how big is this business?

That's how it makes its money, it sells to those people like Amazon for their training of chips, as an example for customers. So, but it is at that data infrastructure world and hyperscalers.

So, what's the market cap for this company? This is a $221 billion market cap, which is a big company by most standards. But when we talk about Nvidia and others in AMD recently, when we talk trillions of dollars for market cap, it is in the billions, but still a large company, $221 billion. So by no means is this a small company, but certainly not in the massive category, the Uber market cap companies like Nvidia, Amazon, Microsoft, Apple, as an example. All right. So let's get into some of the stats here for the business. Again, data infrastructure, chip business, Fabless.

What does it mean to be Fabless? Means they don't have their own manufacturing facilities, as we've talked about some other Fabless companies before, and most recently AMD. Who makes their chips? Taiwan Semiconductor makes the vast majority, and certainly right now, 100 percent of their high-end chips, what they call the two nanometer, for example, chips. All their high-end, three nanometer, two nanometer, 1.4, which is projected nanometer from Taiwan Semiconductor. So all the high-end chips come from Taiwan Semiconductor, which is a whole separate thing that we probably need to check in again on Taiwan Semiconductor, that that is a huge bottleneck for many of these companies, that who controls the high-end of the market on chip manufacturing is Taiwan Semiconductor. Now, a little bit of a side quest here. Part of the reason you can think that SpaceX and Tesla want to do this terafab is to have more control over that output, because a lot of that production is in Taiwan, and a lot of that production is for lots of other customers. And so if you're trying to get a lot of chips, going to be very challenging. Now, they're moving Taiwan Semiconductor, moving some of that into the US, Arizona and other parts of the world, Japan, Germany, but maybe not fast enough really for everybody's demands. All right, back to Marvell just to be off that side quest. But that is something to think about. Taiwan Semiconductor is a really big, obviously company, but an important piece of this ecosystem for AI. And Marvell is using them for all their high-end stuff. They're using Global Foundries for some less high-end things. But really, everything right now for their high-end products are running through Taiwan Semiconductor. They're very important, but they are fabulous, so they do not make those chips themselves. And right now, at least I haven't seen any indication, how they might diversify that with other players in the short term. All right. Now, back to the numbers here describing the business. Certainly, but actually before we move off of that, that's obviously a risk that we have to think about when we think about how concentrated they are on the production side. All right. Now, let's get into the revenue numbers. Remember, market cap $221 billion as of late September when we're recording this for Marvell, ticker MRVL, Mr. VL, if you like. Revenue in the trailing 12 months, so they have a fiscal year of February, so a little bit different, or end of January, excuse me. So a little bit different. And so quarter reported through 801, 2026 trailing 12 months, revenue $9.4 billion. So a good quarter up or good 12 months up 30 percent year over year, so quite good numbers. And if you look at the past couple of years, very good. But let's look even a little further back as we like to do sometimes to get some context of where the business has come from and where it might go. Obviously, for us as business people, analysts, investors, we care all about where the company is going because if we own shares today or if we're buying shares today, remember, educational purposes only, but for our student funds and our students, we have to think about that. Is it in the portfolio? Should it be in the portfolio? All of these questions that we have to think about, it's really about thinking about where the company is going from here. But let's give us some context. There's 2023 revenue, 5.9, and then 2024, 5.5. So actually dipped from 2023 to 2024 Remember, January fiscal year there for the company. Then back to the plus, 5.7. So pretty modest, a little bit less than 5% growth from 2024 to 2025 But then now you're seeing it from 2025 to 26 That's where you see the pickup. Now, that's 42%. The trailing 12 months, we're talking about 30%. The full year projected, which will be 131, 27, 46, nearly 47%.