 Hello and welcome to this post-MWC23 briefing, formerly Mobile World Congress, now MWC. I'm Dave Vellante, and one of the co-hosts of the show with myself and Lisa Martin and joining me today is David Nicholson, CTO extraordinaire, Wharton faculty member and CUBE analyst, Dave. Thanks for making some time. Good to see you in our Palo Alto studio. Yeah, good to see you. Little bicoastal action here. It's great. Yeah, so let's talk about some of the takeaways from MWC. Mine, we're kind of, let's talk about the venue, the event, nearly 100,000 people. They were expecting 80. The venue was awesome. Andrew, show the CUBE set up. It's amazing. It was great. Thanks to our sponsor, anchor sponsor, Dell Technologies. Here's David Nicholson and one of our happy customers. It's our G. Joe Hall, the whole ecosystem participants. We had around 50 sessions, Dave, 100 guests over four days. I mean, just really amazing. Dave, your thoughts on the venue? The venue was amazing. I mean, especially where the FIRA is, the FAIR, I guess that's what it translates to from Catalan. The lines of cabs coming in from all over the city, it just really, it took on a life of its own. In terms of the content that was discussed and the vibe, it's a lot of private 5G. And that was really the gist of the conversation. How do we get there? A lot of perspectives on disaggregation of technology in that space. I love dining at these endless buffet tables of knowledge. So we had a great opportunity. Yeah, so okay, let's talk about telco transformation. I mean, that was the big theme of the show. It's happening, but slowly, but then again, not so slowly when you bring in the cloud. So the large established telcos are trying to protect the past from the future, in my opinion. We heard from the CEOs and senior executives from Orange and Ericsson and Telefonica and others, kind of demand or imply or maybe request, depending on who the executive was, compensation for the way that the over-the-top vendors, the OTT vendors, are bow guarding their network, big boo-hoo from some of the richest monopolies in Europe, I mean, just to give you a sense of the size of the European telcos in terms of revenue, Deutsche Telecom, 73 billion Euro, Telefonica, more than 50 billion, Vodafone, 48 billion, Orange, 40 billion plus, BT, 23 billion, Telecom Italia is almost 19 billion and so forth. And these are monopolies, David. We heard from Greg Peters, who's the co-CEO at Netflix. He fired back. He said, hey, we Netflix, we invest massive amounts, $60 billion in content and maybe the carriers should pay a tax for the content that's driving so many people to their network. That was a fun conversation at the show, wasn't it? Oh yeah, it was. And I always revert to my consumer hat and I'm sitting there thinking, all I want is pizza. Do the cheese manufacturers and the crust manufacturers and the sauce manufacturers really have to fight over all of this? Just deliver me my pizza. But I thought it was great, frankly, that Netflix hit back because I think they have a valid point, they're creating content. Without that content, there wouldn't be the need for some of this network stuff we're talking about. I mean, it's the way I see it. You've got sort of a bifurcated telco world. You've got the big established guys, they're all kind of looking at the cloud, but they're afraid of the cloud. Some of them are partnering with the likes of AWS and Azure and Google. Others are being really aggressive about it. And so that's got to play out. But look, the big conversation in enterprise infrastructure was around the disaggregation of the telco stack. And Andrew, if you can bring up that next slide. And Dave, I want to talk about this. You know, here we're showing the, on the left side, the traditional telco stack, if you will, you know, it's got the everything's integrated, the hardware, the infrastructure, the security, the metrics and telemetry, the OSS, the BSS, which are like the crown jewels. And, you know, while we like to criticize sometimes the telcos, let's face it, they do a good job of not going down. Our phones are always working. The mobile network is pretty darn good, except if you're a heart of Palo Alto or in 128. But that's beside the point. I mean, think about during the pandemic, everything shifted to landlines, or many things did. They didn't miss a beat. But on the right hand side, we're repeating in a way what happened in the 1990s in enterprise tech, where you have the disaggregation or disintegration of the stack. You got infrastructure hardware and cloud services. We've added a little piece in there in Silicon and I want to come back and talk about that. And then you got, you know, infrastructure software, you got cloud services. I think the data layer is going to emerge. And then you've got applications and services on top of that with innovation coming from developers and new workloads. And really, as the notation is here, the focus is on innovation, flexibility, openness, new ways of monetization, because telcos got to monetize 5G. At least that's their mantra. And it's a big opportunity for a lot of different constituents. Your thoughts? Yeah, so just coming to the studio up the 2.80 this morning, it was interesting. I looked over and there was a cellular tower right next to the highway. And right there at the base was the BBU, the shack of gear. And it was kind of fun knowing that, hey, I know what's going on in that shack now. But I always look, I look at these things with a question, with sort of a primary question. The question is, what problem is being solved by a move in tech? Is it a customer problem? Or is it a problem that a vendor might have? And you outlined it perfectly. We get pretty darn good service from the existing model. So what is the impetus for moving to disaggregation? Is it simply because there are vendors that want to sell stuff, that can't sell stuff now? Or are there legitimate reasons to move to that model? I'd like to hear your take on this, but I think that a pretty good case was made that this disaggregation brings with it agility and efficiency and flexibility that just can't exist in the monolithic model. And you and I have both been a veteran of this before. This isn't the first time we've been to this movie. Look at the world of Oracle, Exa, fill in the blank. The idea that if you manage an entire stack in a vertical way, you're gonna get a better value proposition. Those battles will go on forever. But I'd like to get your thoughts on, do you think the case was made that disaggregation is worth the risk? I think, yeah, so thank you for that. I mean, I think that there are some similarities and some differences. I think the differences are, you've got a much broader ecosystem. You know, developers rule the roost. You've got the cloud, which you didn't have back in the 1990s. You've got things like converged infrastructure or engineered systems. You also have a much more competitive semiconductor environment. And that's why on that chart, we sort of added that little tiny layer of semiconductors. And I want to talk about that, but let's look at, let's take open RAN, Dave, a Radio Access Networks. You know, they are highly reliable historically. They don't go down. They're built for resilience. And now open RAN is trying to drive innovation and get costs down. They're not there yet. But everybody, I think, generally agrees that in the fullness of time, all RAN is going to be open. But the open RAN crowd really has to do some work to convince the traditional telcos anyway, that this is a safe bet. You've got companies like Rakuten who are really focused on taking open RAN and making it really resilient. And then you got, on the other end of the spectrum, you've got Dish Networks who's really driving innovation really hard. And I think, you know, both models, there's room for both models because it's a multi-trillion-dollar marketplace. But that open RAN is interesting to me, Dave, because it's so server-intensive and it's going to consume a lot of, you know, old MIPS, right, the old days, a lot of MIPS, but it's going to consume a lot of compute power. So it's got to be, it's got to be, you know, low power. It's got to be efficient. It's got to be great TCO. Not there yet, but we're moving in that direction. Yeah, yeah. And when we talk about MIPS or, you know, the modern equivalent, we tend to focus on CPUs. And this happens to be the season of CPU leaps in terms of speeds and feeds. But that's where the discussion gets more interesting in this open world, where you're talking about everything on that motherboard real estate that surrounds the CPU, memory architecture, PCIe bus architecture, the things that plug into those buses like networking cards and storage controllers, those all become critically important. And if you're going to play in the open ran space, you need to be able to effectively disaggregate the old ran world and then re-aggregate in a way that delivers at least as much reliability and resiliency plus something better, whether it's a combination of better economics or just better economics, but hopefully it's better economics and better plus something else, you know, something that it is greater than the sum of the parts, agility, efficiency, the ability to push new application environments out. All of that is dependent upon all of those piece parts that are in the system. It's not just the CPU. Yeah, so let's talk about private 5G because that was another big conversation. There's a spectrum of no pun intended of 5G, private 5G solutions, a lot of talk about, well, is this going to cannibalize Wi-Fi? Probably not, but there is a van there. We saw, I think we had on theCUBE, we had a small little, I think Dell brought on, a little wireless private 5G box. I was like, hey, I want one of those for my home or my office. What's your take on private 5G, where that fits in the marketplace? And then I want to really dig in under the covers and get into some of the enabling technologies around silicon. Well, this is, this actually reminds me of something. You know, whenever we go into a subject matter like this, I try to gain as much expertise in advance as I possibly can. So I came into the show thinking that I understood this whole discussion around private 5G. It was about after day one, though, that a light bulb really went off and I realized that the private 5G conversation isn't necessarily focused on the big existing telco providers. It's actually forging a net new market for folks outside of that space. So independent in some cases of the large telco providers who aren't going to be refreshing their tech very quickly. So that's, you know, the, I think that's a critical understanding that people have to have is that when we talk private 5G, we're talking about using technology that is in fact being used by the large telco providers but will be rolled out more slowly more quickly. So you take a large enterprise, can leverage things from a private 5G perspective and from an open ran perspective that a large telco provider just simply can't deploy as quickly. Okay, let's go on to the covers a little bit. Dave, there were a number of semiconductor companies at the show, many with very large booths and I want to talk about this a little bit because silicon is a key enabler of that sort of open, disaggregated stack and sometimes it just doesn't get the attention. Qualcomm was there, Broadcom is doing some interesting things, although they didn't have a lot of news, NVIDIA, Intel, AMD, Marvell and a number of others are like contributing to this next generation of telco infrastructure and I want to get you take on what's happening here. You know, back in the day, it was all about the CPU or I remember years ago, the trade pubs they'd asked like Bill Gates, he was like this rock star, what his favorite PC was and he picked the most powerful and latest and greatest Intel based system, the 286 or the 386 or the Pentium or whatever it was and the journalists would swoon. This just seems so dumb today, but the point I'm making is we're moving from a system architectures that have traditionally been CPU centric to a world that's connect centric where the surrounding components are increasingly responsible for gains in performance, price performance, lower cost and just more efficient power consumption, which was a huge topic at the MWC23 and you mentioned some of these components before, but my question is, do you agree with this premise and what are those under the cover components? How does it relate to what's going on in the telecommunication space? Yeah, so I definitely agree with the premise. Increasingly things like NICS, Network Interface Cards moving from 25 to 100 gigabit and beyond make a critical difference specifically in the telecom space. One of the things that was sort of, that was really driven home for us in the interviews that we did was this idea that hardware needs to be designed specifically with the constraints that the telecom world puts upon the vendors that build these systems. Things as simple as the servers going into a rack and you won't have access to the back of that rack. So things have to be custom designed so everything can be pulled out in a modular way in sleds. Take that example of Network Interface Cards in the advances there. You go from 25 gig to 100 gig, you go from more efficient circuitry on that card and the port consolidation, the consolidation, the density increases become the difference between a successful deployment of OpenRAN and an unsuccessful impossible deployment of OpenRAN. Storage controllers, another example. There are instances where storage controllers and onboard rate acceleration makes sense. In other cases, it's simply expansion of the amount of storage that needs to be attached on the back end. All of those things are driven and opened up by latest generation bus architectures that are part of the motherboards that these new generation CPUs plug into. So we love to pull out the AMD or the Intel CPU and wave it around and then plug it back into the board. People often don't ask, well, what's all that other stuff around there? Stuff that companies like Broadcom produces, those components are at least as important as those CPUs. I want to double click on this because I'm breaking analysis. We've done a lot of work covering this space, myself and David Foyer and others. And I want to just do a level set for our audience. Andrew, I wonder if we could bring up the next slide. We're going to compare several semiconductor names and show their performance over the past 12 months. And of course, we're talking stock performance. You can see we're showing Broadcom, is that top blue line and Nvidia, Qualcomm is the yellow, Marvell is the purple and Intel is the orange. Now keep in mind that the NASDAQ is down about 12% for the past 12 months and semis have been hit hard as PC and smartphone demand have tailed off. You got China lockdowns have compounded the problem. So we've gone from the pandemic where we had a shortage of semiconductors which is really severe to now, we're in a semiconductor glut. But Dave, as you can see, Qualcomm, Intel and Marvell have absolutely pounded the bulls. And so really been hurt. But two companies, Broadcom and Nvidia are actually in the green over the past 12 months, which is interesting to me. And the reason it's interesting is because you got Broadcom shelling out $61 billion for VMware, which normally would scare investors. Yet investors are rewarding the company relative to its peers. And we'll see why in a moment. Now Nvidia is a different story. The company's revenue dropped by 20% last quarter year on year. And Andrew, I wonder if you could bring up that chart again. It's crypto business is obviously in the tank along with the Bitcoin and gaming which boomed during the pandemic is not as robust. But Nvidia, you can see here is holding up pretty well. And the reason is because Jensen last quarter did a masterful job of connecting Nvidia to GPT, with GPT washing. But Dave, what's your take on the market and the role of semiconductor manufacturers generally in telco and next generation servers? Yeah, well, I think you just nailed it. I was thinking about everybody chasing after AI. And look, to pick up any trade publication literally and virtually and everybody's racing to associate themselves with AI and saying that they've got something AI enabled about what they do back to pizza. I'm waiting for AI enhanced pizza myself. But no, I think it's really interesting. I think what's shaping up frankly is kind of a clash of Titans around the high ground in AI and high performance computing. And the big players there are gonna be Nvidia and Broadcom. Right now, I see what Broadcom is delivering quietly behind the scenes. I think you mentioned kind of this, talking about Mobile World Congress, MWC as it's now known. Broadcom wasn't there front and center because they're sort of behind the scenes on everything. It's not just the chipsets that enable 5G within mobile devices, but it's the critical components in next gen servers that they produce. So I think they're pretty well positioned as a practical matter from a business perspective to move forward. I personally think that Nvidia is benefiting from a bit more hype than reality at this point. But they all play a critical role that gets overshadowed by the CPU manufacturers and their partners, their partner vendors that integrate systems, whether it's for telecom or anything else. I mean, I get a lot of respect for these semiconductor companies. Nvidia's obviously was a high flyer and is riding the AI wave. Definitely a fan of Qualcomm and others. Intel, I've been very critical of. David Floyer and I, back in 2012, said, wow, this company is peaked and they got to make some moves and they just didn't. But let's dig into this a bit more. Andrew, if you could bring up the last slide. This is a classic Dave Vellante breaking analysis snapshot. I'm not doing a forward looking view here. I'm taking a trailing 12 month TTM look and the purpose is really to compare the financial models and just do a quick snapshot and it'll give you a decent sense. And so you got Broadcom, Qualcomm, Nvidia, Intel. We've added AMD, you can't have, you only have five on the previous chart that I showed. And then Marvell. So the first row is TTM revenue and billions. And you can see Broadcom's at 34 billion, Qualcomm 43, Nvidia 27 billion, Intel 63 billion, AMD 24 billion, mostly coming out of Intel's pocket and Marvell's relatively tiny, but these are large, well-established companies. The next row is the year-on-year revenue growth from the last quarter. And you can see only Broadcom and AMD grew last quarter in the double digits. Qualcomm down 12%, Nvidia down a whopping 20%, Intel down 28% and missed its revenue forecast by more than a half a billion dollars. Now, you see Marvell growing a little bit. Gross margins are more than decent given that these guys are making components. But look at Broadcom. Broadcom's 10 points higher than Nvidia and Qualcomm and it flows right to the operating margin. Broadcom's got a 44% operating margin versus 33% for Qualcomm, which is really good. And 21% for Nvidia, which is also good. And then single digits for the rest. Then look at the market caps. Nvidia's got a $593 billion market cap as of today, this afternoon, midday. With a 22x trailing 12 month revenue multiple. That's like software multiple better. Broadcom, which to me looks to be a stronger business model, has only an 8x revenue multiple but it's still considerably better than the others. Now, again, Dave, this is a backwards looking view. Some of my Wall Street buddies will say, Dave, you got to get the Bloomberg terminal. You got to get the Bloomberg subscription and get the forecast in there. But it does give you a sense of which companies may be overvalued and which may be undervalued. And as I said, Nvidia is really, you know, hopped on the GPT momentum around AI, but Broadcom's business appears to be solid. Now you add VMware to the mix and with the CA acquisition, almost half its business is going to be throwing off software like margins. They guided last quarter, Broadcom, that gross margins are actually increasing. And that's without the VMware contribution. So really interesting company, as you said, they're kind of quiet under the covers, but your thoughts on these names generally and how they'll contribute to the enterprise tech space, but specifically the telco space. Yeah, so I think first and foremost, the A and the I in David, in my David, not your David, mean that I'm AI enabled. So I want to raise, we'll talk about that later. But I think it's interesting when you look at market caps. Market caps are always a combination of execution and speculation. And because of the hype cycle around AI, which a lot of it is warranted, you're gonna see trends where multiples will be higher anytime someone like myself, who is AI enabled as I shared with you, they're gonna benefit from that hype. But frankly, all of these companies are going to be instrumental in the future of AI. But the real question is, what does the future look like when someone like Broadcom tilts that balance so much in the direction of software? Generally speaking, software is considered to be a higher margin opportunity. So you think someone getting more into software would bode well for market cap. And as long as they execute and Broadcom has demonstrated that they are an execution master under Hock Tan and his crew. So I think that this is gonna be really interesting to follow moving forward. I stay away from buying individual stocks cause I'm bad at predicting, but I'm certainly fascinated with following the story. Yeah, I got somebody who's much smarter than I on stock picking, taking care of my stuff. I sleep at night, but what I tell you this day, if I were Michael Dell, I also would have sold VMware for $61 billion to Hock Tan because he's got 40 billion reasons to do that himself in Silver Lake. But if he didn't, I would have spun it in to Dell because that is an awesome asset. It's how Dell restructures, it's how Dell bought EMC, got VMware basically as part of the deal. Was able to go from private to public again and put $40 billion in incremental value back in their pockets after shelling out $4 billion for the EMC transaction. Absolutely phenomenal. But from a strategy standpoint, Dell is now left without a major software component. Broadcom is now picking that up. It's a really interesting dynamic that's going on in the marketplace and one that I'm excited to see play out. Yeah, yeah. Well, I had a front row seat for much of that going back before 2004. I was at EMC when EMC acquired VMware. Someone needs to write a book and I'm going to propose the title of VMware, spelled W-H-E-R-E, right? Yeah, VMware. Kind of tracking the development back at Stanford all the way through the discussions that were happening when that ended up with EMC acquiring them and then everything that's happened in between. It's been absolutely fascinating and they're forging away forward in the era of cloud. It's going to be really, really interesting to watch. So that goes back to this, the question of Silicon and the vendors that are involved and the critical role that they play. The hope, if you're a Broadcom fan, is that this enhances their position in terms of their ability to deliver world-class Silicon to drive things like telecom. Yeah, and I think your point about AI, my prediction is within five years, virtually every job is going to be powered by AI. There's going to be very few exceptions to that. And I do think this conversation that we've been having around, CPU-centric versus connect-centric, I also think alternative processors, you've certainly seen the ascendancy of NVIDIA. I love the ARM model because I watch what companies like Tesla doing and doing their own Silicon. And I think it's very disruptive, particularly at the edge, AI inferencing at the edge. And I think increasingly it's going to be the combinatorial factors of not only the CPU but the other components that you talked about that are really going to drive innovation in the future. Dave, we got to go. Thanks so much for spending some time with me. You got to check out siliconangle.com. We've got tons of articles still flowing from MWC. Check out the cube.net for the MWC23 coverage. The cube was super excited to be there. For Dave Nicholson, this is Dave Vellante and we'll see you next time.