MWC Barcelona 2024 | theCUBE Insights Unpacking Cisco AI with Zeus Kerravala | ZK Research

Clips
News
More from MWC Barcelona 2024

Zeus Kerravala

Founder & Principal Analyst

ZK Research

Analysts dissect Cisco’s challenges and its path toward AI networking

At a time when numerous tech companies are jockeying for a lead position in artificial intelligence, Cisco Systems Inc. has found itself having to play catch-up.Two analysts from theCUBE Research participated in a briefing at MWC Barcelona on Tuesday in which Cisco’s chief executive Chuck Robbins discussed his firm’s challenges. Earlier this month, the company announced plans to trim 5% of its global workforce and issued weaker than expected guidance in its earnings report.“He did admit they’ve lowered expectations two quarters in a row, which he said they’ve never done before,” said Zeus Kerravala (pictured, second from right), guest analyst for theCUBE

play_circle_outline Insider Insights: Analyzing Cisco's Exclusive Event, Chuck's Business Presentation, and Supply Chain Shortages

play_circle_outline Importance of the network in the age of AI

play_circle_outline The significance of investing in infrastructure prior to network support for AI: Insights into Cisco's AI capabilities and operational potential

play_circle_outline Analyzing Cisco, HPE, and Juniper: Networking Capabilities and HPE's Potential to Rival Cisco with AI Turnkey Solution

play_circle_outline Dell's Networking Missed Opportunity: Evaluating the Crucial Value of Silicon versus General-Purpose Processors

Info

theCUBE Insights Unpacking Cisco AI with Zeus Kerravala | ZK Research

Zeus Kerravala

Founder & Principal Analyst ZK Research

search

(bright music)

>> Good evening and welcome back to Mobile World Congress here in beautiful Barcelona, Spain. My name is Savannah Peterson and we are at the end of day two of four days of wall to wall coverage. Very excited to bring you these analyst insights to close out what has been a fantastic Tuesday here in Barcelona. I am joined by Dave Vellante, John Furrier, and Zeus. You are a Cube collective member of VIP Distinguished Cube Alumni. You write for Silicon Angle. You're recently a part of the team. You're one of our posters.

>> Well, I got a punch card, so after so many visits, it is free, so.

>> But we feel very lucky to have you and we're excited that you have our punch card. I know that you and Dave had a very exclusive analyst access.

>> We did

>> To a Cisco event this morning. Tell us a little bit about that.

>> Yeah, that was interesting. So it starts off with Chuck giving the state of the business and he was pretty frank. You know, he said that a lot of what's going on with their business right now was just the fact that, the way I describe it actually is, you know, think of when you go to a grocery store and you're really hungry, you just buy everything you can see, right?

>> Yeah. >> And so, we went through the pandemic.

>> Recklessly >> Where they had supply chain shortages and customers couldn't get anything. And so as soon as the supply chain freed up, that opened up the floodgates and customers bought lots and lots of infrastructure. It's not just Cisco saying that, you know, HPE said that, Dell said that. I've heard Juniper talk about that.

>> So was it Cisco? Is it Cisco event? What was the round table there?

>> Oh, so it was their analyst round table.

>> Cisco systems?

>> With Cisco executives.

>> So when you go to the Cisco booth, very nice booth. Like many of the booth are out there.

>> It's a play by play.

>> You go to the Cisco booth, you check in, and then they walk you back and there's several rooms, maybe, I don't know, dozen rooms.

>> Yeah.

>> We went up the stairs. It's actually quite nice. But the rooms were small, so they brought us downstairs, sat us down in a big room. There were probably like, what, 20, 25 seats?

>> Yeah. >> It was a big sort of--

>> 20 seats.

Square round table. >> Big square. They put Chuck, Jonathan.

>> Zeus: And Jeff Sharritts, the head of sales.

>> Jeff Sharritts, the head of sales and--

>> Zeus: And Mark Patterson.

>> Mark.

>> Yeah. >> Which is a--

>> Zeus: Chuck, chief of staff.

>> Chief of staff, right. And then Zeus and I, of course, had the prime seats 'cause we got there early, so.

>> Savannah: Smart, gentlemen, well done.

>> Well, it's no accident, right?

>> Definitely no accident for sure.

>> But Chuck did start with the state of the business. And one of the things that surprised me a little bit, Dave, I don't know if you picked up on this, was he did admit they've lowered expectations two quarters in a row, which he said they've never done before. And they had heard the noise that a slowdown was coming from customers, said they were having a hard time digesting all the product they bought, but then the impact of that was greater than they realized. And that's why they had to reduce the numbers the second time. And Cisco's a very disciplined company. They understand pipeline and sales, you know, better than anybody out there. And so the fact that they got caught off guard with that, to me says a lot changed in that three month period, right?

>> Yeah. So he said, "Two quarters in a row, we've lowered guidance." And that's not a good thing for a CEO to do.

>> That's what I've heard of for Cisco.

>> No. >> Yeah.

>> So, guys, breaking news. So, we know Cisco had a 5% riff reduction in force, which is across the board cut. One hour ago, tech layoffs continue at Cisco plans to cut another 700 jobs.

>> Yeah. >> The word leaked out after that. Somebody told me that they were going to do it again. You know, they have to, right? They have to do this. And so, you know, lowering guidance twice, I mean, let's go back. I mean, you guys both remember the .Dot-com boom.

>> Yeah.

>> In February of 2000, Cisco hit over a $300 billion market cap. They were the most valuable company in the world.

>> Zeus: You're right. Over 800 actually.

>> Yeah. Over 800?

>> Yeah, they were talking about them being the first trillion dollar company

for a while in 2000. >> Wow, wow.

>> Wow. >> So, yeah, yeah.

>> So, okay. >> They lost more market cap between then and 2001 than any company ever held market down.

>> Dave: So they've never gotten back to that level.

>> Savannah: Wow, that's a crazy data point.

>> Yeah, yeah.

>> Stock price was in the se mid-'70s up there, you know, something like that. And they've never gotten back there. They've gotten close, but never got back there. So, anyway, a lot of people are comparing what's going on with Nvidia to what happened with Cisco saying, "Oh, it's picks and shovels." Personally, I think it's different. I don't know what you think.

>> Well, my thinking is Cisco had the monopoly on routing. Everyone knows that they're having a routing system. It's hard to rip out and replace. So competitors had to have a huge bar to get and win against Cisco because they were so nested in there. Switching costs were extremely high and costly. So, all the competitors just fell by the wayside or found niches that they could come in on an angle and then John Chambers went on an acquisition run.

>> Yeah. >> And then, they just had piece parts. So to me, I think what's different between Nvidia and Cisco is Nvidia actually did the work to build software that works across their core asset. In this case, the GPUs. Cisco never really did anything outside of their core asset routing. They had some nice businesses that stood alone, but like switches were a gift. But that's the top of rack, that's racks and routers. So you had routers and switches and racks. That's networking. They dominated. Outside of that, they just bought businesses and they were trying to cobble them together. And they rationalized it. I mean, they didn't go under, but they didn't really knock it out of the park. And I think that was probably what couldn't get them back. Dave, if I had to go back and look at that. Zeus, let me get a little of your perspective, but--

>> Yeah, so one of the interesting things about Cisco is that they have transformed that company a lot. There's 51% ARR now. I think the 30 something percent software. And actually one of the things I liked that Chuck said today was, on the earnings call, they're not going to really talk about, he doesn't want to talk about transformation anymore. Transformation's done, just talk about the business. And what's going to be interesting to see is once they close Splunk, they're going to be well over 50% ARR. Now, if you look at there are multiple right now.

>> Yeah. >> They're trading at about 4x, which is where the hardware companies are. Software companies that large of they are, trade from 8x to, I don't know, infinity. Right? And so once they close Splunk.

>> Yeah. >> What I'd like to see is, are investors going to give 'em more of a software multiple or are they going to keep 'em in the hardware penalty box?

>> I think they're going to do--

>> An interesting question.

>> They got to do a CUDA moment. They need a CUDA moment for Cisco, which means they have to take their core platform, get the data out of it. I think they got to look at the generative AI and saying, "We have to kind of make this a true platform with observability to get in and get the AI action." We heard from Dell's CTO and Telecom achieve this morning saying, "What is the impact of AI and telco?" He said, "It goes to the device." Cisco's got edge, data. So the question is--

>> Zeus: Well, doesn't Splunk actually give a headstart?

>> Dave: Wait, wait, wait. Hold on, hold on, hold on, hold on.

>> I think Splunk is a real piece of the puzzle.

>> Cisco is not going to have a CUDA moment and here's why. They can't. Okay. The CUDA moment for Nvidia happened around late 2000s, I want to say 2008, where they basically said to the street, "We're going to spend a boatload of money developing this software." And essentially, we don't care about our stock price. And the thing tanked. There's no way Cisco can do that. There's no way that--

>> They're kind of tanking now. So, was that--

>> No. >> I mean.

>> That's what Microsoft did too, right?

>> Get destroyed.

>> Actually let that company burn to the ground before he built it back up.

>> But the difference with Microsoft is they had a software estate that threw off so much cash. Now Cisco throws off $15 billion of cash flow a year. But to do that, the CEO would get fired. Now maybe Chuck will initiate that. It's probably, I don't know, I was surprised that--

>> Their software estate office a little different comparable. Apples to oranges, but still platform that gave them leverage. I think my view on Cisco is they have a leverage platform with networking and they got market share numbers that are awesome. So they have to kind of, when I say CUDA, I mean, they need a software layer that actually is going to be more than just window dressing.

>> Zeus: Well, that's what they're--

>> Dave: Well, yeah, they're doing this.

>> That's the new security cloud and the networking cloud. They've consolidated all those products. They created a single backend for security. One for networking. And that's what they have to monetize, right? And because right now, what they do is they sell hardware and then you buy the cloud after. 'Cause they need to get customers thinking, the cloud products first.

>> It's the ecosystem, yeah.

>> So notwithstanding that it's apples oranges software company versus hardware company, it's more like Microsoft, to your point, Zeus, 'cause they went sideways for years and then they really need a Nadella moment.

>> Yeah. >> Right? Somebody to come in or even a Gelsinger moment, even though I think he's got big challenges.

>> That's implying that Chuck is going to be replaced.

>> Well, he is. I mean, actually it was surprising to me. I thought Chuck was like 52 years old. He's like 64.

>> Is he really?

>> He's 64 years old.

>> I guess all the golf he plays. Keeps him in shape.

>> Well, he looks fantastic, but somebody pointed out to me, I said, "He's not retiring. The guy's young." Looked it up, he's 64 years old.

>> John: Has he announced his retirement?

>> No, but--

>> John: So you're expect--

>> He's been there eight years.

>> Yeah. >> Anyway, my point is--

>> I think he's got a few left.

>> My point is, yeah, maybe a few. But my point is as networking goes, Cisco goes.

>> Yeah. >> Yeah.

>> And so for them to have, forget the CUDA like moment, but it's going to be a long transformation. They're trying to do that with Splunk. Patrick Moorhead said today, "Well, you could be like Cloudera."

And I was like, "Huh?" >> Huh?

>> Yeah, >> That made no sense to me. I'd be like, "Cloudera, what are you talking about? You're going to start doing database, data management?" No, but with Splunk, they could have a big data platform. The question is--

>> So, here's the question. Does the CUDA moment come to them?

Right? >> Cloudera. They're a customer of theirs. That's why.

>> So instead of chasing the CUDA moment, it could come to them. We live in a world today that's network centric. Everything we do from mobile devices, cloud, IOT, connected cars, all runs over the network, right? So the quality of your network--

>> Yes. >> Really determines how's your in-store experiences, how's the experience in the stadium is, how hospitals work, how schools work. And I think from here, they need to convince the world that the center of your IT strategy is the network that'll create that value prop. And I think it'll bring the CUDA moment to them.

>> Yeah, I agree with that. I think that's the unification.

>> That's a leap from there, right?

>> Well, they got to do the work. That's a unification, that's a value proposition. That's a disruptive enabler. It has to enable value and that spits off revenue. So to me, you know, they have to use the observability trend that's being retooled. And clearly from this show, my takeaway, Cisco looks like a wounded animal right now. Okay, they look like the way they're posturing in their--

>> That's a little harsh.

>> Yeah. >> Well, I mean, well, they just laid off 4,000 people, another 700.

>> Zeus: Amazon's been laying off people. Apple's been laying off people.

>> I mean, everyone's laid off people in the last couple of months.

>> Okay. >> I think the thing. Go ahead, please.

>> So I think what you're all touching on here, in the details of this, is the winners of the next few years are going to be the ecosystem players. It's the conversations that we've been having. It's not just going to be this, you know, yes, network is a part of it, but I think it's going to be who's pulling in the biggest collaborators. And that's one of the things that Nvidia does incredibly well.

>> Yeah. >> They're an ecosystem partner with pretty much everyone here.

>> Well, that's because the GPUs become the center of the world.

>> Yeah. >> Right?

>> In this particular supply chain moment.

>> Right. >> And this particular AI moment.

>> And so, if the network does pivot that, who's got a bigger ecosystem than Cisco on the networking side?

>> Right. >> Yeah.

>> Well, exactly why there's still a lot of opportunity.

>> I mean, here's-- >> I've been covering Cisco since 1995, been close to the company in the Valley on the east coast, watching their technology, did routing tables at one point as an ST in my earlier days. They're a great company that their IP on the routing, switching, networking. They have basically a monopoly in my mind. They run networks. That is such a massive--

>> They do well. >> They have a monopoly.

>> I think they're as good as a monopoly 'cause hard to rip 'em out. How do you replace Cisco? No one. Name the people that replaced Cisco. You got some niches that they get established in certain markets. But Cisco has such a great nested position.

>> Dave: What about HPE Juniper?

>> I think they're affordable competitor now that they're merged together, you know?

>> By the way, I forgot to adjust for a share count. It was 546 billion at their peak.

>> Okay. >> Sorry. But yeah, and you're right. Everybody said it could surpass a trillion, they surpassed Microsoft.

>> Yeah. >> But here's what's missing. My opinion from Cisco. They have a full AI stack starting with Silicon, the networking, security, observability. They've got collaboration applications. They've got AI all throughout the stack, but they don't communicate an AI strategy. They're getting no tailwind for that.

>> Yeah. >> Totally agree

with that, Dave. >> It's very bespoke.

>> Totally agree with that.

>> Yeah, they need to pull that. I think that's their CUDA moment. I think there's an opportunity in there for them. It would pull them together.

>> They should have an AI czar. Have they hit you up?

>> We have great retrieval on the keyboard. Well, the question about Juniper is interesting.

>> But okay, they have good AI to help really with AI ops, letting their customer use the products more efficiently. They don't really sell AI, you know what I mean? Like--

>> Dave: Well, they claim they sell networks for AI.

>> Right?

>> So, the bottom of the stack. And you're right, they don't sell AI.

>> But I would argue though-- >> But are there ways they could monetize it?

>> I mean, I would argue though. We used to always say that every company is a tech company, now every company is an AI company. So even if you're not selling it, you're supporting it or you're providing solutions in that space.

>> Yeah. >> And if you're not telling that story, how are you going to attract other customers?

>> Yeah. >> And I think you brought up a great point, Dave. We don't hear a lot from them.

>> Yeah, Robins did say though, they had 25 billion in pipeline for AI. Right, that should close the next couple years. So that's going to be interesting.

>> So what was that? I asked them that question.

>> But I'm curious about that, that could be

a made up number. >> Like, can you double click

on that? >> Yeah.

>> I talked about the waves and the transitions, how they're really difficult. The old and the new.

>> So that's the network that touches AI systems, right?

>> Dave: So that's Silicon One?

>> Oh, okay. >> Yeah, it's all Silicon One stuff, right, and--

>> Okay, so the other question I was asking him, basically implying that, so Nvidia is sucking up all the CapEx.

>> That was an amazing question, I thought.

>> And he didn't really bite on it, but so Nvidia last quarter, this is Charles Fitz, hit 50%, their revenue hit 50% of the CapEx spend of all the three big three hyperscalers, which is an amazing stat. Now that doesn't mean that they--

>> Zeus: That's unbelievable.

>> That was all the CapEx there, but, yeah. But a big chunk of the spend went to the hyperscalers. I think they said over half of their revenue went to hyperscalers. So you're talking about, you know, 11, 12 billion, just sucked up by hyperscalers this past quarter. My point is, are they delaying some of the network purchases? And he sort of said, "No, no, they're still buying." But what's that giant pipeline then in Silicon One? Why can't they service it?

>> Well, I think, you know, customers, I saw Chuck's interview with CNBC on when he was at Davos too. And he actually made a comment, he didn't really think customers were all that far along in AI. And I think that's, you know, that data point where he talked about pipeline is probably it. So, there's plans to do it. You got to buy a lot of the infrastructure first and then you bring the network to support it. And so--

>> One of the things I agreed with that he said is a lot of customers are saying, "We're not going to spend, we're keeping it for AI. We don't know what we're going to spend it on. We know it's AI."

>> Yeah,. >> But we're not going to spend

it on other. >> Well, they did that with cloud and things like that in the past.

>> The question is, what does AI for Cisco mean? I mean, like I said, the AI that they could harvest for their own benefit for a platform is different than what they offer as a solution. So, where's the value proposition for the customer? Well, there's certainly, I'll make the argument that Cisco should tap AI for their own operational platform 'cause they have a lot of data. And then two, what's the pivot to the enablement side to say, "Hey, we're going to be a disruptive enabler for you, the customer." So, like routing and switching, that was an enablement opportunity to connect people, devices, and campuses. What's the AI value prop?

>> We invited Chuck on to talk about this. You know, he probably never got the invite, but Antonio, I thought and Juniper.

>> Yeah. >> Step Up.

>> Rami, Step Up. Michael Dell came on and talked about their AI strategy. To me, it's much more clear.

>> Yeah. >> So I'd love to have. We invited Jonathan on as well. Hopefully, he'll come on tomorrow.

>> Yeah, but when I say wounded animal, I'm a little bit harsh. I will admit that, Cisco.

>> I understand what you're saying. They're taking shots right now.

>> Their posture, they're bunkering in, right? They're like, okay, they got to heal. They got a massive layoff. Their press and their analysts motions are kind of settled. Subdued a little bit, you know?

>> They had 80,000 employees. So you're talking 8% so.

>> But to your point though. >> I'm not impressed with them.

>> But to Zeus's point, they lowered guidance and then they had a lower guidance again, what, like 2.4 billion?

>> John: Yeah.

>> I mean, that's do big downturn. You lower guidance to a point where you know you can hit it or beat it.

>> Right. >> Yeah.

>> Especially with a company as venerable as Cisco.

>> There's an old saying on Wall Street, "The first disappointment is rarely the last." And that was the case here.

>> Yeah.

>> Yeah, I do think, you know, what I brought up in the Rami interview yesterday though, that AI for networking needs to be in place before you can do networking for AI. And I think that's really, it'll be fascinating to see which one of these networking companies actually is able to capitalize on that. Because I do believe.

>> Likewise. >> Chuck, that he's got the big pipeline. Jay Reese said the same with Arista. They're going to have massive pipeline for networking for AI. But if you can't do AI in the network, I'm not so sure you should be deploying networks that support AI.

>> Why is that Pipeline not, it's just 'cause they can't deliver? I mean, NVIDIA's got a huge pipeline too. But they. (indistinct)

>> Because I think GPUs-- >> Different product.

>> John: So if you look historically.

>> 'Cause they can't get GPUs.

>> No, Compute has always led networking and sale, right? So, people move to branch offices, then you buy all the network infrastructure to support it. People move to the cloud, then you buy all the network infrastructure to support it. So if they're buying GPUs now.

>> Dave: Compute leads networking.

>> Yes. >> Is what you're saying. (people talking over each other drowns out speaker) And to your point, they got a monopoly.

>> Yes. >> You feel like--

>> I love Cisco's value. I'll always love that company. They have love.

>> Well, that's the big question. Do they get more than a 4x multiple after Splunk closes?

>> We'll see how they can get to the software side. Rather than patching together acquisitions and kind of like putting a little close the curtain behind everything and saying, "They kind of work together." But do they really work together? So that's the question.

>> But Jay Reese's getting an AI multiple.

>> Yes. >> Right?

Mist. >> Nobody else--

>> Mist got an AI multiple.

>> Zeus: Mist did, but Juniper didn't.

>> Okay, but where would they be without Mist? They wouldn't be getting 14 billion.

>> John: They got multiple sucked into their evaluation.

>> But they wouldn't be getting 14 billion from HPE.

>> No, they would not. >> Without that. So, what if they pay what they pay for Mist? 400 million, roughly 430?

>> Yeah. >> Yeah.

>> How much of that 14 billion is Mist in your estimation?

>> John: A lot.

>> Oh, a good, at least a third.

>> A third, 30 to 40%? >> Yeah, yeah.

>> Dave: That's what I thought.

>> All right, so question for you guys, since you brought up, Dave, about Juniper and HPE, can they compete with Cisco? All right, now that we had Antonio, the CEO of HPE and the CEO of Juniper on, and since you guys had a little side briefing with the CEO of Cisco, what is the opportunity for HPE? Can they compete with Cisco?

I mean, I think Aruba. >> I think--

>> Aruba couldn't hang on their own but clearly.

>> Well, I think--

>> Aruba's doing pretty well.

>> Yeah, I think the opportunity here is, you know, we've been talking about, I know you're a big fan of engineered systems, right?

>> Yeah. >> Because the world disaggregate everything and nobody can put the stuff back together. And HPE is as full stack of company as anybody. Storage servers. I mean, now they have all the networking stuff.

>> Absolutely. >> They've got AI and if they can apply Mist to Compute, they could deliver a full turnkey AI system that's storage, service, networking, AI, all that put together, right? And I think that's HPE's opportunity to compete with Cisco. Now Cisco's got the same assets minus the storage, I suppose.

>> What seems all based on networking, 'cause Juniper has a lot of service providers, but do they have the enterprise penetration that--

>> They have the enterprise Compute penetration. Cisco's got network and you need both to make AI work. So here comes what, where, who's going to win the battle?

>> HPE has enterprise Compute penetration, right? So the interesting thing to me is if you're Dell, you're looking at this going, "Hmm, maybe we missed an opportunity here to network."

>> No, no.

>> Dell's messed their network strategy.

>> They're done.

>> Yeah.

>> What do you mean no networking?

>> I think they missed the networking thing. I agree with you.

>> Yeah, I mean, but--

>> Their Compute server business is going to transform into clusters.

And Michael-- >> Explain that.

>> So Michael Dell smiled yesterday when I brought up the whole systems revolution and then Etab was on earlier.

>> Yeah. >> And he said, "No. The system's going to be clustered." So, that's like a server. So sales on, if you're HP and Dell, you're in the systems business. You build PCs and servers. Basically, these micro clouds are going to be clustered on premise cluster clouds. And so, Dell's going to mop that up. I think Dell's going to mop that Compute business up and the GPUs will shift into shifting between GPUs and CPUs and these engineered systems.

>> I think Dell's the problem.

>> Dell's Achilles heel is no network.

>> Exactly, here's the problem with--

>> Ethernet. >> No, wait, here's the problem

with that. >> But they don't have that.

>> Hold on, here's the problem with that is servers is a 30% gross margin business. Maybe 35%.

Why? >> At best.

>> Why? Because Intel and now Nvidia are sucking up all the gross margin.

>> Savannah: Yep.

>> Networks is a 65% gross margin business. So, that's why I like--

>> That's a great point.

>> Antonio's move saying, "We're now a networking company."

>> And they have Silicon too. So they got Silicon as well.

>> Smart. >> Yep.

>> Well, it was interesting to hear Rami talk about, you know, on the one hand, you know how valuable it is to have Silicon, but I don't think he was really high on Silicon One. So, it's like--

>> Of course. >> Merchant silicon.

>> He doesn't work there.

>> I don't know if you understand why people have such a hard time understanding the value in network silicon. Like, the analogy I use is, if you look, why is the GPU a thing? Well, because general purpose processors couldn't do graphics, right? And what's Silicon One is, it's the network equivalent to that. You can't run, I mean, there's so much stuff you do with a network processor that general purpose processors can't do.

>> Yep.

>> You're better off building your own. I mean, Fortinet does that with security, right? And so they get a big price to see that.

>> I think if you have the volume--

>> System on Chip, yeah.

>> Is a great business model too.

>> I mean, they have the volume.

>> Yeah, if you have the volume, it makes perfect sense to me to design Silicon.

>> So merchant silicon versus custom chips day. What was your takeaway from your, we've had a lot of call way conversations. Guys, what's your take on, you know, 'cause the custom silicon and systems on the chip stuff is good purpose-built stuff. But--

>> Merchant silicon has to be differentiated. AMD's differentiated 'cause they're beating Intel. Broadcom is differentiated 'cause they're focused on connectivity. Obviously, Nvidia is differentiated. Qualcomm is executing on a differentiation strategy. So if you have a differentiated strategy and you have wafer volume, it makes sense. I do think that Cisco One has a differentiated strategy. They've got design shops. You're saying they've got volume.

>> Yeah, I mean, so look at the size of their network business.

>> Dave: Well, to me, it makes sense.

>> Yeah. >> Amazon, Google, Microsoft, they've all got wafer volume capability. So it makes sense for them to design their own silicon.

>> Okay.

>> I actually think, I think Dell should think about designing its own silicon, but they've said, "That's not for us."

>> So, Juniper. What's their alternative to the Silicon One?

>> Well, they've got their own silicon.

>> Slingshot. >> Yeah, yeah.

>> Dave: Right?

>> John: How's that volume?

>> HPE Slingshot. >> Well, the volume is the size of their install base, so.

>> It's tiny, HPE Slingshot's tiny, but it's very specialized around high performance computing. So, maybe they can get away with that, you know?

>> The reality is we're going to need a lot of different silicon solutions to make it through this next AI leap across verticals in tech as a whole.

>> John: Go to SiliconANGLE.com. You got all the action.

>> Yeah, I mean, there's startup players too. There's companies like Rock. It's not just going to be the old school players.

>> And that statement actually is why Intel is in the poor position that they're in.

>> Well, I think too, I think too. A lot of people rightly say exactly what you just said, there's going to be a lot of silicon diversity.

>> Yeah. >> And a lot of specialization. But I would say this, very few are going to be able to get monopoly like margins and I think Nvidia will and I think they'll be able to sustain them for quite some time.

>> I think that's a really brilliant note to end on. Otherwise, we could end up talking about Silicon and Cisco for the next two hours here.

>> Dave: I know this was kind of a Cisco--

>> We missed a whole Amazon web services conversation. We had Dr. Ishwar Parulkar, chief technologist. Interesting. Amazon's pushing that. But we can get that tomorrow on the kickoff.

Maybe. >> Yeah. We can start hot and fresh after this wonderfully spicy analyst insights right now.

>> But we did, if I may, we did have some awesome guests on today and yesterday.

>> We really did, Dave.

>> I think if, you know, Antonio.

>> Caitlin. >> Kicked it off.

>> Yeah, Caitlin Halferty, who's the chief data officer at Erickson. Rami was guest of the day. No offense. Well, you actually were on that so that was good. Rami was amazing. Michael Dell was heading out to the airport, so he was kind of in and out. But we really appreciate him coming on and he gave us some great insights and more to come.

>> Yeah, yeah. >> Yeah, we had an honest teach, we've had some really great stories on. And we are only halfway through our coverage, folks, from our four days of coverage here live in Barcelona, Spain at Mobile World Congress. My name's Savannah Peterson. And you're watching theCube, the leading source for enterprise tech. (bright music)

theCUBE Insights Unpacking Cisco AI with Zeus Kerravala | ZK Research

search

(bright music)

>> Good evening and welcome back to Mobile World Congress here in beautiful Barcelona, Spain. My name is Savannah Peterson and we are at the end of day two of four days of wall to wall coverage. Very excited to bring you these analyst insights to close out what has been a fantastic Tuesday here in Barcelona. I am joined by Dave Vellante, John Furrier, and Zeus. You are a Cube collective member of VIP Distinguished Cube Alumni. You write for Silicon Angle. You're recently a part of the team. You're one of our posters.

>> Well, I got a punch card, so after so many visits, it is free, so.

>> But we feel very lucky to have you and we're excited that you have our punch card. I know that you and Dave had a very exclusive analyst access.

>> We did

>> To a Cisco event this morning. Tell us a little bit about that.

>> Yeah, that was interesting. So it starts off with Chuck giving the state of the business and he was pretty frank. You know, he said that a lot of what's going on with their business right now was just the fact that, the way I describe it actually is, you know, think of when you go to a grocery store and you're really hungry, you just buy everything you can see, right?

>> Yeah. >> And so, we went through the pandemic.

>> Recklessly >> Where they had supply chain shortages and customers couldn't get anything. And so as soon as the supply chain freed up, that opened up the floodgates and customers bought lots and lots of infrastructure. It's not just Cisco saying that, you know, HPE said that, Dell said that. I've heard Juniper talk about that.

>> So was it Cisco? Is it Cisco event? What was the round table there?

>> Oh, so it was their analyst round table.

>> Cisco systems?

>> With Cisco executives.

>> So when you go to the Cisco booth, very nice booth. Like many of the booth are out there.

>> It's a play by play.

>> You go to the Cisco booth, you check in, and then they walk you back and there's several rooms, maybe, I don't know, dozen rooms.

>> Yeah.

>> We went up the stairs. It's actually quite nice. But the rooms were small, so they brought us downstairs, sat us down in a big room. There were probably like, what, 20, 25 seats?

>> Yeah. >> It was a big sort of--

>> 20 seats.

Square round table. >> Big square. They put Chuck, Jonathan.

>> Zeus: And Jeff Sharritts, the head of sales.

>> Jeff Sharritts, the head of sales and--

>> Zeus: And Mark Patterson.

>> Mark.

>> Yeah. >> Which is a--

>> Zeus: Chuck, chief of staff.

>> Chief of staff, right. And then Zeus and I, of course, had the prime seats 'cause we got there early, so.

>> Savannah: Smart, gentlemen, well done.

>> Well, it's no accident, right?

>> Definitely no accident for sure.

>> But Chuck did start with the state of the business. And one of the things that surprised me a little bit, Dave, I don't know if you picked up on this, was he did admit they've lowered expectations two quarters in a row, which he said they've never done before. And they had heard the noise that a slowdown was coming from customers, said they were having a hard time digesting all the product they bought, but then the impact of that was greater than they realized. And that's why they had to reduce the numbers the second time. And Cisco's a very disciplined company. They understand pipeline and sales, you know, better than anybody out there. And so the fact that they got caught off guard with that, to me says a lot changed in that three month period, right?

>> Yeah. So he said, "Two quarters in a row, we've lowered guidance." And that's not a good thing for a CEO to do.

>> That's what I've heard of for Cisco.

>> No. >> Yeah.

>> So, guys, breaking news. So, we know Cisco had a 5% riff reduction in force, which is across the board cut. One hour ago, tech layoffs continue at Cisco plans to cut another 700 jobs.

>> Yeah. >> The word leaked out after that. Somebody told me that they were going to do it again. You know, they have to, right? They have to do this. And so, you know, lowering guidance twice, I mean, let's go back. I mean, you guys both remember the .Dot-com boom.

>> Yeah.

>> In February of 2000, Cisco hit over a $300 billion market cap. They were the most valuable company in the world.

>> Zeus: You're right. Over 800 actually.

>> Yeah. Over 800?

>> Yeah, they were talking about them being the first trillion dollar company

for a while in 2000. >> Wow, wow.

>> Wow. >> So, yeah, yeah.

>> So, okay. >> They lost more market cap between then and 2001 than any company ever held market down.

>> Dave: So they've never gotten back to that level.

>> Savannah: Wow, that's a crazy data point.

>> Yeah, yeah.

>> Stock price was in the se mid-'70s up there, you know, something like that. And they've never gotten back there. They've gotten close, but never got back there. So, anyway, a lot of people are comparing what's going on with Nvidia to what happened with Cisco saying, "Oh, it's picks and shovels." Personally, I think it's different. I don't know what you think.

>> Well, my thinking is Cisco had the monopoly on routing. Everyone knows that they're having a routing system. It's hard to rip out and replace. So competitors had to have a huge bar to get and win against Cisco because they were so nested in there. Switching costs were extremely high and costly. So, all the competitors just fell by the wayside or found niches that they could come in on an angle and then John Chambers went on an acquisition run.

>> Yeah. >> And then, they just had piece parts. So to me, I think what's different between Nvidia and Cisco is Nvidia actually did the work to build software that works across their core asset. In this case, the GPUs. Cisco never really did anything outside of their core asset routing. They had some nice businesses that stood alone, but like switches were a gift. But that's the top of rack, that's racks and routers. So you had routers and switches and racks. That's networking. They dominated. Outside of that, they just bought businesses and they were trying to cobble them together. And they rationalized it. I mean, they didn't go under, but they didn't really knock it out of the park. And I think that was probably what couldn't get them back. Dave, if I had to go back and look at that. Zeus, let me get a little of your perspective, but--

>> Yeah, so one of the interesting things about Cisco is that they have transformed that company a lot. There's 51% ARR now. I think the 30 something percent software. And actually one of the things I liked that Chuck said today was, on the earnings call, they're not going to really talk about, he doesn't want to talk about transformation anymore. Transformation's done, just talk about the business. And what's going to be interesting to see is once they close Splunk, they're going to be well over 50% ARR. Now, if you look at there are multiple right now.

>> Yeah. >> They're trading at about 4x, which is where the hardware companies are. Software companies that large of they are, trade from 8x to, I don't know, infinity. Right? And so once they close Splunk.

>> Yeah. >> What I'd like to see is, are investors going to give 'em more of a software multiple or are they going to keep 'em in the hardware penalty box?

>> I think they're going to do--

>> An interesting question.

>> They got to do a CUDA moment. They need a CUDA moment for Cisco, which means they have to take their core platform, get the data out of it. I think they got to look at the generative AI and saying, "We have to kind of make this a true platform with observability to get in and get the AI action." We heard from Dell's CTO and Telecom achieve this morning saying, "What is the impact of AI and telco?" He said, "It goes to the device." Cisco's got edge, data. So the question is--

>> Zeus: Well, doesn't Splunk actually give a headstart?

>> Dave: Wait, wait, wait. Hold on, hold on, hold on, hold on.

>> I think Splunk is a real piece of the puzzle.

>> Cisco is not going to have a CUDA moment and here's why. They can't. Okay. The CUDA moment for Nvidia happened around late 2000s, I want to say 2008, where they basically said to the street, "We're going to spend a boatload of money developing this software." And essentially, we don't care about our stock price. And the thing tanked. There's no way Cisco can do that. There's no way that--

>> They're kind of tanking now. So, was that--

>> No. >> I mean.

>> That's what Microsoft did too, right?

>> Get destroyed.

>> Actually let that company burn to the ground before he built it back up.

>> But the difference with Microsoft is they had a software estate that threw off so much cash. Now Cisco throws off $15 billion of cash flow a year. But to do that, the CEO would get fired. Now maybe Chuck will initiate that. It's probably, I don't know, I was surprised that--

>> Their software estate office a little different comparable. Apples to oranges, but still platform that gave them leverage. I think my view on Cisco is they have a leverage platform with networking and they got market share numbers that are awesome. So they have to kind of, when I say CUDA, I mean, they need a software layer that actually is going to be more than just window dressing.

>> Zeus: Well, that's what they're--

>> Dave: Well, yeah, they're doing this.

>> That's the new security cloud and the networking cloud. They've consolidated all those products. They created a single backend for security. One for networking. And that's what they have to monetize, right? And because right now, what they do is they sell hardware and then you buy the cloud after. 'Cause they need to get customers thinking, the cloud products first.

>> It's the ecosystem, yeah.

>> So notwithstanding that it's apples oranges software company versus hardware company, it's more like Microsoft, to your point, Zeus, 'cause they went sideways for years and then they really need a Nadella moment.

>> Yeah. >> Right? Somebody to come in or even a Gelsinger moment, even though I think he's got big challenges.

>> That's implying that Chuck is going to be replaced.

>> Well, he is. I mean, actually it was surprising to me. I thought Chuck was like 52 years old. He's like 64.

>> Is he really?

>> He's 64 years old.

>> I guess all the golf he plays. Keeps him in shape.

>> Well, he looks fantastic, but somebody pointed out to me, I said, "He's not retiring. The guy's young." Looked it up, he's 64 years old.

>> John: Has he announced his retirement?

>> No, but--

>> John: So you're expect--

>> He's been there eight years.

>> Yeah. >> Anyway, my point is--

>> I think he's got a few left.

>> My point is, yeah, maybe a few. But my point is as networking goes, Cisco goes.

>> Yeah. >> Yeah.

>> And so for them to have, forget the CUDA like moment, but it's going to be a long transformation. They're trying to do that with Splunk. Patrick Moorhead said today, "Well, you could be like Cloudera."

And I was like, "Huh?" >> Huh?

>> Yeah, >> That made no sense to me. I'd be like, "Cloudera, what are you talking about? You're going to start doing database, data management?" No, but with Splunk, they could have a big data platform. The question is--

>> So, here's the question. Does the CUDA moment come to them?

Right? >> Cloudera. They're a customer of theirs. That's why.

>> So instead of chasing the CUDA moment, it could come to them. We live in a world today that's network centric. Everything we do from mobile devices, cloud, IOT, connected cars, all runs over the network, right? So the quality of your network--

>> Yes. >> Really determines how's your in-store experiences, how's the experience in the stadium is, how hospitals work, how schools work. And I think from here, they need to convince the world that the center of your IT strategy is the network that'll create that value prop. And I think it'll bring the CUDA moment to them.

>> Yeah, I agree with that. I think that's the unification.

>> That's a leap from there, right?

>> Well, they got to do the work. That's a unification, that's a value proposition. That's a disruptive enabler. It has to enable value and that spits off revenue. So to me, you know, they have to use the observability trend that's being retooled. And clearly from this show, my takeaway, Cisco looks like a wounded animal right now. Okay, they look like the way they're posturing in their--

>> That's a little harsh.

>> Yeah. >> Well, I mean, well, they just laid off 4,000 people, another 700.

>> Zeus: Amazon's been laying off people. Apple's been laying off people.

>> I mean, everyone's laid off people in the last couple of months.

>> Okay. >> I think the thing. Go ahead, please.

>> So I think what you're all touching on here, in the details of this, is the winners of the next few years are going to be the ecosystem players. It's the conversations that we've been having. It's not just going to be this, you know, yes, network is a part of it, but I think it's going to be who's pulling in the biggest collaborators. And that's one of the things that Nvidia does incredibly well.

>> Yeah. >> They're an ecosystem partner with pretty much everyone here.

>> Well, that's because the GPUs become the center of the world.

>> Yeah. >> Right?

>> In this particular supply chain moment.

>> Right. >> And this particular AI moment.

>> And so, if the network does pivot that, who's got a bigger ecosystem than Cisco on the networking side?

>> Right. >> Yeah.

>> Well, exactly why there's still a lot of opportunity.

>> I mean, here's-- >> I've been covering Cisco since 1995, been close to the company in the Valley on the east coast, watching their technology, did routing tables at one point as an ST in my earlier days. They're a great company that their IP on the routing, switching, networking. They have basically a monopoly in my mind. They run networks. That is such a massive--

>> They do well. >> They have a monopoly.

>> I think they're as good as a monopoly 'cause hard to rip 'em out. How do you replace Cisco? No one. Name the people that replaced Cisco. You got some niches that they get established in certain markets. But Cisco has such a great nested position.

>> Dave: What about HPE Juniper?

>> I think they're affordable competitor now that they're merged together, you know?

>> By the way, I forgot to adjust for a share count. It was 546 billion at their peak.

>> Okay. >> Sorry. But yeah, and you're right. Everybody said it could surpass a trillion, they surpassed Microsoft.

>> Yeah. >> But here's what's missing. My opinion from Cisco. They have a full AI stack starting with Silicon, the networking, security, observability. They've got collaboration applications. They've got AI all throughout the stack, but they don't communicate an AI strategy. They're getting no tailwind for that.

>> Yeah. >> Totally agree

with that, Dave. >> It's very bespoke.

>> Totally agree with that.

>> Yeah, they need to pull that. I think that's their CUDA moment. I think there's an opportunity in there for them. It would pull them together.

>> They should have an AI czar. Have they hit you up?

>> We have great retrieval on the keyboard. Well, the question about Juniper is interesting.

>> But okay, they have good AI to help really with AI ops, letting their customer use the products more efficiently. They don't really sell AI, you know what I mean? Like--

>> Dave: Well, they claim they sell networks for AI.

>> Right?

>> So, the bottom of the stack. And you're right, they don't sell AI.

>> But I would argue though-- >> But are there ways they could monetize it?

>> I mean, I would argue though. We used to always say that every company is a tech company, now every company is an AI company. So even if you're not selling it, you're supporting it or you're providing solutions in that space.

>> Yeah. >> And if you're not telling that story, how are you going to attract other customers?

>> Yeah. >> And I think you brought up a great point, Dave. We don't hear a lot from them.

>> Yeah, Robins did say though, they had 25 billion in pipeline for AI. Right, that should close the next couple years. So that's going to be interesting.

>> So what was that? I asked them that question.

>> But I'm curious about that, that could be

a made up number. >> Like, can you double click

on that? >> Yeah.

>> I talked about the waves and the transitions, how they're really difficult. The old and the new.

>> So that's the network that touches AI systems, right?

>> Dave: So that's Silicon One?

>> Oh, okay. >> Yeah, it's all Silicon One stuff, right, and--

>> Okay, so the other question I was asking him, basically implying that, so Nvidia is sucking up all the CapEx.

>> That was an amazing question, I thought.

>> And he didn't really bite on it, but so Nvidia last quarter, this is Charles Fitz, hit 50%, their revenue hit 50% of the CapEx spend of all the three big three hyperscalers, which is an amazing stat. Now that doesn't mean that they--

>> Zeus: That's unbelievable.

>> That was all the CapEx there, but, yeah. But a big chunk of the spend went to the hyperscalers. I think they said over half of their revenue went to hyperscalers. So you're talking about, you know, 11, 12 billion, just sucked up by hyperscalers this past quarter. My point is, are they delaying some of the network purchases? And he sort of said, "No, no, they're still buying." But what's that giant pipeline then in Silicon One? Why can't they service it?

>> Well, I think, you know, customers, I saw Chuck's interview with CNBC on when he was at Davos too. And he actually made a comment, he didn't really think customers were all that far along in AI. And I think that's, you know, that data point where he talked about pipeline is probably it. So, there's plans to do it. You got to buy a lot of the infrastructure first and then you bring the network to support it. And so--

>> One of the things I agreed with that he said is a lot of customers are saying, "We're not going to spend, we're keeping it for AI. We don't know what we're going to spend it on. We know it's AI."

>> Yeah,. >> But we're not going to spend

it on other. >> Well, they did that with cloud and things like that in the past.

>> The question is, what does AI for Cisco mean? I mean, like I said, the AI that they could harvest for their own benefit for a platform is different than what they offer as a solution. So, where's the value proposition for the customer? Well, there's certainly, I'll make the argument that Cisco should tap AI for their own operational platform 'cause they have a lot of data. And then two, what's the pivot to the enablement side to say, "Hey, we're going to be a disruptive enabler for you, the customer." So, like routing and switching, that was an enablement opportunity to connect people, devices, and campuses. What's the AI value prop?

>> We invited Chuck on to talk about this. You know, he probably never got the invite, but Antonio, I thought and Juniper.

>> Yeah. >> Step Up.

>> Rami, Step Up. Michael Dell came on and talked about their AI strategy. To me, it's much more clear.

>> Yeah. >> So I'd love to have. We invited Jonathan on as well. Hopefully, he'll come on tomorrow.

>> Yeah, but when I say wounded animal, I'm a little bit harsh. I will admit that, Cisco.

>> I understand what you're saying. They're taking shots right now.

>> Their posture, they're bunkering in, right? They're like, okay, they got to heal. They got a massive layoff. Their press and their analysts motions are kind of settled. Subdued a little bit, you know?

>> They had 80,000 employees. So you're talking 8% so.

>> But to your point though. >> I'm not impressed with them.

>> But to Zeus's point, they lowered guidance and then they had a lower guidance again, what, like 2.4 billion?

>> John: Yeah.

>> I mean, that's do big downturn. You lower guidance to a point where you know you can hit it or beat it.

>> Right. >> Yeah.

>> Especially with a company as venerable as Cisco.

>> There's an old saying on Wall Street, "The first disappointment is rarely the last." And that was the case here.

>> Yeah.

>> Yeah, I do think, you know, what I brought up in the Rami interview yesterday though, that AI for networking needs to be in place before you can do networking for AI. And I think that's really, it'll be fascinating to see which one of these networking companies actually is able to capitalize on that. Because I do believe.

>> Likewise. >> Chuck, that he's got the big pipeline. Jay Reese said the same with Arista. They're going to have massive pipeline for networking for AI. But if you can't do AI in the network, I'm not so sure you should be deploying networks that support AI.

>> Why is that Pipeline not, it's just 'cause they can't deliver? I mean, NVIDIA's got a huge pipeline too. But they. (indistinct)

>> Because I think GPUs-- >> Different product.

>> John: So if you look historically.

>> 'Cause they can't get GPUs.

>> No, Compute has always led networking and sale, right? So, people move to branch offices, then you buy all the network infrastructure to support it. People move to the cloud, then you buy all the network infrastructure to support it. So if they're buying GPUs now.

>> Dave: Compute leads networking.

>> Yes. >> Is what you're saying. (people talking over each other drowns out speaker) And to your point, they got a monopoly.

>> Yes. >> You feel like--

>> I love Cisco's value. I'll always love that company. They have love.

>> Well, that's the big question. Do they get more than a 4x multiple after Splunk closes?

>> We'll see how they can get to the software side. Rather than patching together acquisitions and kind of like putting a little close the curtain behind everything and saying, "They kind of work together." But do they really work together? So that's the question.

>> But Jay Reese's getting an AI multiple.

>> Yes. >> Right?

Mist. >> Nobody else--

>> Mist got an AI multiple.

>> Zeus: Mist did, but Juniper didn't.

>> Okay, but where would they be without Mist? They wouldn't be getting 14 billion.

>> John: They got multiple sucked into their evaluation.

>> But they wouldn't be getting 14 billion from HPE.

>> No, they would not. >> Without that. So, what if they pay what they pay for Mist? 400 million, roughly 430?

>> Yeah. >> Yeah.

>> How much of that 14 billion is Mist in your estimation?

>> John: A lot.

>> Oh, a good, at least a third.

>> A third, 30 to 40%? >> Yeah, yeah.

>> Dave: That's what I thought.

>> All right, so question for you guys, since you brought up, Dave, about Juniper and HPE, can they compete with Cisco? All right, now that we had Antonio, the CEO of HPE and the CEO of Juniper on, and since you guys had a little side briefing with the CEO of Cisco, what is the opportunity for HPE? Can they compete with Cisco?

I mean, I think Aruba. >> I think--

>> Aruba couldn't hang on their own but clearly.

>> Well, I think--

>> Aruba's doing pretty well.

>> Yeah, I think the opportunity here is, you know, we've been talking about, I know you're a big fan of engineered systems, right?

>> Yeah. >> Because the world disaggregate everything and nobody can put the stuff back together. And HPE is as full stack of company as anybody. Storage servers. I mean, now they have all the networking stuff.

>> Absolutely. >> They've got AI and if they can apply Mist to Compute, they could deliver a full turnkey AI system that's storage, service, networking, AI, all that put together, right? And I think that's HPE's opportunity to compete with Cisco. Now Cisco's got the same assets minus the storage, I suppose.

>> What seems all based on networking, 'cause Juniper has a lot of service providers, but do they have the enterprise penetration that--

>> They have the enterprise Compute penetration. Cisco's got network and you need both to make AI work. So here comes what, where, who's going to win the battle?

>> HPE has enterprise Compute penetration, right? So the interesting thing to me is if you're Dell, you're looking at this going, "Hmm, maybe we missed an opportunity here to network."

>> No, no.

>> Dell's messed their network strategy.

>> They're done.

>> Yeah.

>> What do you mean no networking?

>> I think they missed the networking thing. I agree with you.

>> Yeah, I mean, but--

>> Their Compute server business is going to transform into clusters.

And Michael-- >> Explain that.

>> So Michael Dell smiled yesterday when I brought up the whole systems revolution and then Etab was on earlier.

>> Yeah. >> And he said, "No. The system's going to be clustered." So, that's like a server. So sales on, if you're HP and Dell, you're in the systems business. You build PCs and servers. Basically, these micro clouds are going to be clustered on premise cluster clouds. And so, Dell's going to mop that up. I think Dell's going to mop that Compute business up and the GPUs will shift into shifting between GPUs and CPUs and these engineered systems.

>> I think Dell's the problem.

>> Dell's Achilles heel is no network.

>> Exactly, here's the problem with--

>> Ethernet. >> No, wait, here's the problem

with that. >> But they don't have that.

>> Hold on, here's the problem with that is servers is a 30% gross margin business. Maybe 35%.

Why? >> At best.

>> Why? Because Intel and now Nvidia are sucking up all the gross margin.

>> Savannah: Yep.

>> Networks is a 65% gross margin business. So, that's why I like--

>> That's a great point.

>> Antonio's move saying, "We're now a networking company."

>> And they have Silicon too. So they got Silicon as well.

>> Smart. >> Yep.

>> Well, it was interesting to hear Rami talk about, you know, on the one hand, you know how valuable it is to have Silicon, but I don't think he was really high on Silicon One. So, it's like--

>> Of course. >> Merchant silicon.

>> He doesn't work there.

>> I don't know if you understand why people have such a hard time understanding the value in network silicon. Like, the analogy I use is, if you look, why is the GPU a thing? Well, because general purpose processors couldn't do graphics, right? And what's Silicon One is, it's the network equivalent to that. You can't run, I mean, there's so much stuff you do with a network processor that general purpose processors can't do.

>> Yep.

>> You're better off building your own. I mean, Fortinet does that with security, right? And so they get a big price to see that.

>> I think if you have the volume--

>> System on Chip, yeah.

>> Is a great business model too.

>> I mean, they have the volume.

>> Yeah, if you have the volume, it makes perfect sense to me to design Silicon.

>> So merchant silicon versus custom chips day. What was your takeaway from your, we've had a lot of call way conversations. Guys, what's your take on, you know, 'cause the custom silicon and systems on the chip stuff is good purpose-built stuff. But--

>> Merchant silicon has to be differentiated. AMD's differentiated 'cause they're beating Intel. Broadcom is differentiated 'cause they're focused on connectivity. Obviously, Nvidia is differentiated. Qualcomm is executing on a differentiation strategy. So if you have a differentiated strategy and you have wafer volume, it makes sense. I do think that Cisco One has a differentiated strategy. They've got design shops. You're saying they've got volume.

>> Yeah, I mean, so look at the size of their network business.

>> Dave: Well, to me, it makes sense.

>> Yeah. >> Amazon, Google, Microsoft, they've all got wafer volume capability. So it makes sense for them to design their own silicon.

>> Okay.

>> I actually think, I think Dell should think about designing its own silicon, but they've said, "That's not for us."

>> So, Juniper. What's their alternative to the Silicon One?

>> Well, they've got their own silicon.

>> Slingshot. >> Yeah, yeah.

>> Dave: Right?

>> John: How's that volume?

>> HPE Slingshot. >> Well, the volume is the size of their install base, so.

>> It's tiny, HPE Slingshot's tiny, but it's very specialized around high performance computing. So, maybe they can get away with that, you know?

>> The reality is we're going to need a lot of different silicon solutions to make it through this next AI leap across verticals in tech as a whole.

>> John: Go to SiliconANGLE.com. You got all the action.

>> Yeah, I mean, there's startup players too. There's companies like Rock. It's not just going to be the old school players.

>> And that statement actually is why Intel is in the poor position that they're in.

>> Well, I think too, I think too. A lot of people rightly say exactly what you just said, there's going to be a lot of silicon diversity.

>> Yeah. >> And a lot of specialization. But I would say this, very few are going to be able to get monopoly like margins and I think Nvidia will and I think they'll be able to sustain them for quite some time.

>> I think that's a really brilliant note to end on. Otherwise, we could end up talking about Silicon and Cisco for the next two hours here.

>> Dave: I know this was kind of a Cisco--

>> We missed a whole Amazon web services conversation. We had Dr. Ishwar Parulkar, chief technologist. Interesting. Amazon's pushing that. But we can get that tomorrow on the kickoff.

Maybe. >> Yeah. We can start hot and fresh after this wonderfully spicy analyst insights right now.

>> But we did, if I may, we did have some awesome guests on today and yesterday.

>> We really did, Dave.

>> I think if, you know, Antonio.

>> Caitlin. >> Kicked it off.

>> Yeah, Caitlin Halferty, who's the chief data officer at Erickson. Rami was guest of the day. No offense. Well, you actually were on that so that was good. Rami was amazing. Michael Dell was heading out to the airport, so he was kind of in and out. But we really appreciate him coming on and he gave us some great insights and more to come.

>> Yeah, yeah. >> Yeah, we had an honest teach, we've had some really great stories on. And we are only halfway through our coverage, folks, from our four days of coverage here live in Barcelona, Spain at Mobile World Congress. My name's Savannah Peterson. And you're watching theCube, the leading source for enterprise tech. (bright music)