The Six Five team discusses Marvell IA Day.
If you are interested in watching the full episode you can check it out here.
Disclaimer: The Six Five Webcast is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors and we ask that you do not treat us as such.
Transcript:
Daniel Newman: At Marvell, and you and I did the double duty. That’s what we do here. We run our research firms and then we do a night job, as you like to say. Our night job is PG. We are on video, but it’s talking in technology. Yep, there you go. Got it right. And so we sat down with the President, Chief Operations Officer of Marvell, rising star, Chris Koopmans.
Chris has kind of been like a six-month visitor to the Six Five. He did our summit, one of our main keynotes. He visited us on the ground about six months before that. Now six months later, we were doing the recap. Look, there’s a lot there, Pat. The data infrastructure story around AI is palpable. Marvell was probably the second company that got real credit around AI because they were the second company after NVIDIA that was able to announce a meaningful revenue number. Now, again, we’re talking different stratospheres in terms of the size of the numbers, but Marvell has seen it’s on a quarter to quarter run rate. They went from about 200 million a year of AI infrastructure and data infrastructure to 200 million a quarter over a short period of time.
During that period of time, Marvell has been able to really communicate the story that yes, part of the AI story is all about compute, but the other part of the story is all about moving the data. And so Marvell has got critical infrastructure that is necessary and that the OEMs are going to build with, their modules that are going to be required to be able to move data light and electronic. And that’s where Marvell’s growth story has come from. Having said that, the company didn’t just talk about data infrastructure. They did talk a little bit about the rest of their business across the auto and 5G, but it was a heavy day of AI. It was a heavy day of AI and I think they were calling it accelerated infrastructure. Accelerated data infrastructure was the word they were using to align with accelerated computing. The other thing that really hit home for me, Pat… And of course watch the Six Five video. That’ll come out shortly. You’ll see us sit down with Chris.
The other thing that really came out to me too is this is a company that’s become a really reputed and required partner to the ecosystem for cloud optimized silicon. We’re going to see accelerators coming from Marvell, I believe in the coming year. They’re going to have their first AI accelerator, cloud optimized silicon. They’re a partner to many OEMs and many that are developing silicon. They’ve kind of been a critical one-stop shop white glove service. That has been something that the company’s been able to lean on meaningfully to gain business. Now this is an area we’re growing competition. You’ve got companies like arm, they’re announcing their own IP plus white glove, but Marvell was there early, they’ve been there often and they’ve been very successful. This is also a company that has seen a lot of its business parts struggle. And so while AI has been up, everything else has kind of been sideways and down.
If we’re turning a corner, that’s going to be really good for Marvell, but otherwise they’re going to need to see that AI revenue start really quickly to offset other slower parts of their business, Pat. But it was a good day overall. Lots of good conversations, good demonstrations. What’d you see?
Patrick Moorhead: That was great analysis, Dan. The way that Marvell hasn’t had the growth overall, as I’m sure they would like, but what I always like to do is look at self-inflicted wounds versus market wounds. The reality is that carrier automotive and enterprise is down. It’s down for everybody. Every single semiconductor company we talk to. Now, that one niche, the Hyperscaler AI niche is growing, just like you said. Marvell has had some exceptional growth in there, and they participate primarily on the network side and they have some custom ASIC for AI that is ready to hit the market here. And what do you do as a company if you’re in a position like that? You just trudge on and you move forward, because you know the market is going to come back. It’s like the ocean. The ocean goes out and the ocean comes in, and you need to be prepared when it comes in that you’re the best prepared to do that.
And the company I believe is investing in core intellectual property, building blocks, and also creating solutions, complete solutions to be able to address these core markets competitively. Two main announcements they made at the show. First of all, OCTEON 10 processors, two new ones. These ones are for network equipment folks, firewall folks. It really is a giant SOC, right? We call them DPUs. It contains our Neoverse 2 cores. It includes some ASICs on there to do acceleration. Two new ones out. Those are primarily used in the enterprise, and also I’ll call tier two hyperscalers. Then they came out with quite frankly what I think is their biggest play long-term, and that’s upping the game with optical DSPs. If nothing else that the industry agrees with, is generative AI brings this explosion of data like we’ve never seen before. The reality is if you don’t connect the racks, if you don’t connect the fleets of racks, if you don’t connect the fleets of racks together and you don’t connect data centers between each other in the fastest, lowest latency, lowest power possible, then you’re not going to have growth.
And as we talked with Chris, and I think you and I have known this, but it really is the trifecta. You can’t have a fiddle or a crab architecture where you’ve got a giant compute accelerator and not the right connectivity and not the right memory and storage. All of these things need to come together. Two new solutions came out. One was called Perseus, and this is optical short reach. Dan, you and I did a lab tour of AWS that we’re allowed to talk about.
Daniel Newman: Yes.
Patrick Moorhead: It’s public. What Perseus does is it replaces passive copper cables with active optical cables. So distance are 5 meters to 500, and they brought out this other one that’s called SpikoGen 2 that is 10 kilometers. 10 kilometers. Why would you need an optical solution, a DSP, a PAM-4 optical DSP to do that? It’s connecting data centers together or connecting you with the backhaul of the internet. This is clearly focused on the cloud operators. Those are the three announcements. Talk a little bit about strategy, but please be sure to check out our conversation with chief, our operating officer, Chris Koopmans.
Author Information
Daniel is the CEO of The Futurum Group. Living his life at the intersection of people and technology, Daniel works with the world’s largest technology brands exploring Digital Transformation and how it is influencing the enterprise.
From the leading edge of AI to global technology policy, Daniel makes the connections between business, people and tech that are required for companies to benefit most from their technology investments. Daniel is a top 5 globally ranked industry analyst and his ideas are regularly cited or shared in television appearances by CNBC, Bloomberg, Wall Street Journal and hundreds of other sites around the world.
A 7x Best-Selling Author including his most recent book “Human/Machine.” Daniel is also a Forbes and MarketWatch (Dow Jones) contributor.
An MBA and Former Graduate Adjunct Faculty, Daniel is an Austin Texas transplant after 40 years in Chicago. His speaking takes him around the world each year as he shares his vision of the role technology will play in our future.
