 Hello everyone, welcome to theCUBE, live here in Palo Alto, California and in Boston, Massachusetts, I'm John Furrier with Dave Vellante for a special coverage emergency CUBE and CUBE research analysis on the big enterprise tech news. Today we're going to delve into the major development where ULA Packard Enterprise HPE has announced the acquisition of Juniper Networks, a leader in AI native networks in an all cash deal valued at approximately 14 billion. This strategic move is set to reshape the landscape of the cloud and AI native networking market enhancing HPE's edge to cloud capabilities. Dave, we're going to explore the deals, industry, technology, product and financial aspects for Juniper. Shareholds will receive $40 a share, representing significant premium. Dave, first I want to walk through with you the research piece of it as the chief research officer and CUBE co-host for 13 years. Heading up the new CUBE research with Wikibon, formerly Wikibon with Rob, Stretch A and Shelley Kramer. Dave, it's been interesting news to unpack here and the reaction from the community we're going to have right after this will be telling this is a big deal. Yeah, you know, I mean, Antonio Neary has been aggressive but conservative with acquisitions since he took over. What I mean by that is he's pretty much made sure that the acquisitions are going to give good return, done a lot of tuck ins. He didn't do the Aruba acquisition but he was certainly there at the time. So this is a big, big number for HPE to absorb. HPE was doing really well in networking and I, you know, as we'll show in some of the data, you know, Juniper was kind of having to figure out what to do next, hitting a little bit of a wall from a growth standpoint but I wonder if we could bring up the data from ETR Andrew, the first slide here. And it shows the net score or spending momentum on the vertical axis and the presence or pervasiveness in the dataset. This is a survey of 1700 IT decision makers. That's on the horizontal axis, that presence in the dataset. Anything at or near that 40% mark with that red line is considered highly elevated. So breaking this down a little bit, you see Cisco's the big whale. Everyone wants a piece of their business. They're the king of networking. They got a large business as you can see in the X dimension. They also own Maraki, the simplified sort of management solution which you can see has very, very strong momentum in the market as does Aruba which is HPE's networking crown jewel. Now the ETR data shows it also shows HPE which is probably just customers saying, yeah, we use HPE networking. Could be the legacy three-com asset. And you see Juniper, which is kind of stuck with a net score in the low single digits. It's revenue was basically flattish last year. And so it's got a security business. It's got some AI assets that HPE can leverage and you put these three data points together and it doubles HPE's networking business, John, as you said, to around 10 billion. It gives it a formidable asset to keep penetrating the market and in going after Cisco, you can see a risk is in there too, extreme networks. Risk has got great momentum, but obviously much smaller. Yeah, we're going to have our panel come in and dissect the impact. But I looked at that graph and it's clear, you know, Cisco, obviously the big player, but if you look at the market, Dave, and you look at the technology synergy between HPE and Juniper in this deal, particularly how their combined portfolios can put innovation back at the center of HPE strategy, specifically in AI and cloud-native environments. This acquisition is expected to double HPE's networking business, creating a formal position play against Cisco and others. What is your take on and what data do you have that looks at the broader market implications, specifically the impact to the numbers? Can you share the data on what the combination looks like? Okay, it's going to double the business. What does that mean financially? Because a lot of people are questioning the math here and I said I love this deal, the math looks great. I mean, networking is the hottest area that nobody's talking about in technology. Yeah, so we've put together some of a quick financial snapshot, Andrew, if you'd bring that up, of some of the players that we showed earlier. So what we show here is trailing 12-month revenue, last quarter's growth rate, which is kind of overstated for Juniper at double digits, trailing 12-month gross margin, TTM operating margin, market cap, revenue multiple, head count and revenue per employee. And the takeaway here, as John, as you just mentioned, Cisco's the big player. They got a $60 billion business that everyone wants a piece of. And you can see how the market rewards Arista's growth rate. They got a 14x revenue multiple and its margin profile is similar to Cisco. So high margin, high growth for Arista Extreme, you got kind of a niche player. But if Andrew, if you go to the next slide and you isolate on HPE, Juniper and Cisco, you can see why this deal makes financial sense, in my opinion, John. HPE's picking up a quality $5 billion plus revenue asset. It's going to have okay growth higher or comparable to HPE's gross margin overall and better operating margin. This should be a creative, I would think, within a year. And HPE is undoubtedly going to find some go-to-market and service synergies as well as leverage Juniper's AI, its security assets. So my bottom line is, to me, John, this makes sense from a financial perspective and is only going to bolster the quality of HPE's earnings. Can you explain what does it mean again? Because I think the numbers speak to us. If you go back to the original slide, Andrew, with all the sector players, because Arista's got great numbers, Dave. Dave, this analysis with Arista, why isn't Arista, I mean, their numbers look fantastic. Why, I mean, obviously HPE and Juniper combined, I can see how you put it in Cisco, but what does this slide tell you about the overall networking industry? Arista's got great metrics. Look at those numbers. Unbelievable. I mean, look at the revenue multiple. Look at how the street rewards the execution that Jay Shree, well, Jay Shree is obviously a machine. She's like a business hero of ours, John. She's been at theCUBE a number of times. She's an amazing operator. But look at the asset that they've built. They're around the same, from a trailing 12 month standpoint, around the same size as Juniper. But look at the revenue multiple and the valuation that they're getting. So they're running at 44% growth last quarter. And they got a market cap that's 78 billion. I mean, that is amazing execution. A very well-run company. Their revenue per employee is very strong. So, you know, the market will reward, still rewards growth despite all this focus on interest rates and valuation. Maybe Arista could have bought Juniper. I mean. Yeah, I'm not sure it would need to. I mean, in a way, you know, Juniper, look it, for HPE, Juniper lifts the quality of HPE's financials, whereas Arista, you know, it would actually dilute them. So I think from HPE's standpoint, this makes a lot of sense. Well, both CEOs at Juniper and HPE, both CUBE alumnus, we've sat down with them many times, Dave, and had a chat. What I found interesting is that, remember, we were really bullish on Aruba, but over the past few years, there's been a lot of industry discussion around some product gaps. Okay, so I asked them directly this morning on the analyst's call what those gaps were and did they have a position in Silicon? Well, Antonio lit up like a tree, Christmas tree, like, oh yeah, we've got all this great stuff. And Rami has started his career in Silicon. I had that story in the CUBE. So the question's going to be is, where does HPE compete? And with the combination of Juniper, because, you know, HPE has had a portfolio of networking, okay? Juniper's had the world-class rounding and switching as their roots and they have a lot of service provider customers, network connectivity, network connectivity and application service has been the core of their business. But remember, they saw the fabric early with software and had contrail that they bought. They were early with contrail on multi-cloud. 2018, Dave, they were really pushing hard on multi-cloud because Juniper had a lot of the cloud players as customers. So, you know, I wonder if Juniper was too early here with software and multi-cloud because contrail was hot. So, you know, now HPE and with that software pedigree and Juniper West was started out in 1998, go back to the free BSD, but then, you know, NetScreen was bought by Juniper around 2008. They brought their security. So Junos was upgraded around 2009. So Juniper's got a lot of jewels in the chest there. Definitely, it's got a lot of, it's got a really strong patent portfolio. It's got a good product portfolio. The other thing, I think you're onto something with the silicon. Remember, HPE's got silicon shops going back to pre-split. They've got the Slingshot ASIC, which is out of the high performance computing Cray business. I'll give another sideline, which is amazing. HPE bought Aruba. I want to say it was 2015, John, 2.7 billion. So what an amazing acquisition. That's been, and the other thing is, you know, they really didn't talk a lot about GreenLake. And if you think about HPE's as-a-service portfolio, they got Compute, obviously. They got that down with their server business. They got Aruba. Now they're adding in Juniper. They got storage, all they're missing in storage is object store. That's kind of a big miss with as-a-service, but they will fill that hole. They'll have the entire lower layer of the substrate nailed. And they're ahead in that as-a-service. Now, they probably don't talk about it so much because Wall Street really doesn't care about that, even though I think they should, because I think it's going to throw off better margins in the future, because as-a-service margins should be better. But, you know, I think it strengthens HPE's story there across the basic infrastructure portfolio. Now they can start tucking in, bringing in networking, sorry, in security and moving up the stack in data, improving on top of their Esmerelle assets. So, you know, Antonio's making moves. I like it. I know a lot of people were criticizing it, but I think it makes some sense. I want to hear what the panel has to think. Well, in summary, I love the deal. I think if you look at Juniper's hidden, I would say hidden jewels, I would say they're not really hidden, but not really recognized was. Their experience with cloud, I think it's significant. Their networking and connectivity chops were going to play well at the edge. I think HPE gets a great asset with Juniper. They now have a real play in the service provider market. You know, Mobile World Congress, MWC's coming up, Dave. You know, we're going to be out in Barcelona with theCUBE there, and we'll continue to bring more expertise and analysis to the table. Again, this emergency cube was really just to set the table on the news. You know, $40 a share, $14 billion. You look at packets acquiring in an all cash deal, Juniper Networks, a leader in AI networking. And remember that, you know, the whole hype here is AI, and all the analysts are like, oh, it's AI AI. You know, I get the missed AI angle here, and that's one of the core things we're talking about in the launch. But I think outside of missed AI, which by the way, we reported on theCUBE years ago, AWS used that re-invent. A lot of people have been using missed at their events. The wireless technology has been proving to be great. And I think that's a gap we see a lot with Aruba. Missed aside, Dave, they got the chops. Routing, switching as their roots, network connectivity application services, core competency, software fabric started way early. I mean, Junos was open source. Again, like I said, in 1998, upgraded with the acquisition of Netscreen. If you remember the days back then, you know, Junos was a free BSD open source program. Netscreen increased that value to Juniper around 2009, and with their kind of security play there. You know, Netscreen, one of the best acquisitions Juniper's ever did. It was a significant, you know, excellent portfolio. So I'm bullish on it. I just want to get the take on the numbers. You think it's good. You think it's good. Yeah, well, I think it's good for both sides, because where was Juniper going to go? I mean, Juniper had a, you know, TAM expansion challenge. They really kind of, the stock's been sideways for a number of years. I mean, they were obviously a hot company back in the dot com days, as it was every networking company. But, you know, with their Juniper, I mean, where were they going to go from here? So I think the marriage of HPE and Juniper makes a lot of sense. HPE with its services chops and its GreenLake capabilities and it's go to market and its distribution channel. I think it's a good move for both companies. I'm a little concerned about, you know, the balance sheet impact is an all cash deal. They're going to, you know, take some cash out of their, their till, take on some debt. And I'm not crazy about that, but that's all right. But you think the numbers look, you think the numbers look good. If you go back to that slide again, I think that shows the cash positions, market cap, earnings, that really kind of hits the mark from your perspective. My big takeaway is it just improves the quality of HPE's earnings. This is not dilutive. This is a creative to HPE's earnings. You know, HPE, think about HPE and Dell. Dell kept its laptop business. HPE spun it off with HPE. So that, that throws off good revenues for Dell, but it drags down the margins. So HPE should be a higher margin business. That was the whole intent of that spin out, higher margin, you know, better growth. And they've been struggling to grow. And so they've got to make moves like this to continue to grow. But to me, this improves the overall quality of HPE's earnings. So from that standpoint, I like it. Let's get into a quick, before we get to the panel, I want to just quickly, what are your open questions that you're going to look at? I mean, to me, I'm looking at the product portfolio overlap, Antonio Neary went and asked that question. He said, oh, there's no overlap. I was kind of rolling my eyes. Yeah, yeah. Okay, really. We'll debate that. The simple overlap. The easy overlap will have a good perspective on our panel on that. Now they asked about the Silicon because as we've been reporting, specifically at reinventing the HPC event of supercomputing 23, the role of the chips and networking and other chips, IO specifically, this is kind of like a motherboard problem, but it's not the motherboard. It's like, what's the architecture of AI in the future? And what's coming up over and over again is that the network component, and we saw that at Nitro aspect for AWS, the role of Nitro is playing in their architecture. If you look at what HP could do here with networking and Silicon, to bring that into the compute GPU and the TPU side of the business, it's an opportunity to recreate some new configurations at the hardware level, chip level. I think this is something I'm going to be watching very closely. It's the Silicon angle, so to speak. We're going to watch that very closely. That's my take on terms of the questions and of course the team consolidation, how fast can you get that done? I think the product overlaps not a big deal, but there is an issue, and then what's going to be the Silicon? How is that going to render itself as a value? I think that the piece two is moving up the stack. I mean, if GreenLake is HPE's cloud, which it is just like Apex is Dell's cloud, they've got to have that substrate, that as-a-service substrate. I think HPE is ahead of Dell in that regard and has done some good work there, but they've got to keep adding services. You think about how Amazon just kept adding services like crazy, now I'm not suggesting that HPE needs to do that, but they've got to have the baseline which they're doing. They've announced LLMs as a service, they've got supercomputing as a service, they've got some Esmerell assets in data. I'd like to see more in data management. They keep tucking in up the stack. This is a big move, it's going to take some time to absorb it. I'm really curious to see how it integrates into the whole GreenLake because that's their NorthStar. That's their operating NorthStar, everything gets GreenLaked. And so I'm curious to see how they're able to execute on that front. Great data you shared, data is really timely for this emergency cube. Appreciate that, and the cube research team delivering again. The moment's noticed, quick data, quick analysis, great job on that. Put a quick plug-in for the cube research. I know you've got a lot going on, you're building out, formerly Wikibon now kind of rebooting as a new name, the cube research. Give a quick plug for the cube research for the folks watching this might want to hear more. Yeah, we're rebranding the cube research, formerly Wikibon. And I think as always, the Wikibon ethos has always been to try to identify new markets, to create new markets. I mean, we've done it with SuperCloud, we've done it with software-led infrastructure. We were really early on with the arms ascendancy and there are half a dozen examples of markets that we created or early on into and we're going to continue that trend right now. One of our big themes is the six data platform. What's the next generation of data management look like and in real time, we're working with our practitioner audience, we're working with the cube collective, which is a group of really strong independent analysts and thought leaders. So we're just combining that insight with the cube and its amplification capabilities, John, as well as our partnership with ETR, which is the best enterprise tech survey business, you know, in the industry. Well, 13 years of the cube, everyone loves the knowledge, they love the learning, they love the insights, they love the amplification, love the promotion of the fresh voices, the experts. We're going to keep doing it. Again, a lot more going on. On this emergency pod, we're going to have a reaction right now with a panel of experts to dissect the technology synergies between HPE and Juniper, particularly how the combined portfolios lead to innovation in this new AI and cloud-native environments. This acquisition is expected to double the networking business, creating a formal player. We're going to get the analyst perspective on the implications, the broader market implications, including the impact on the customers and their partners in the channel. And then again, does this position HPE to accelerate on the macro trends around AI? We're going to get it all now. Up next with an expert panel, we'll be right back with that panel after the short break. Hello, and welcome back to our special coverage and emergency cube on the big enterprise news. We're delving into the development where HPE Hewlett Packard Enterprise has acquired Juniper Networks for $40 billion, a $40 a share for $14 billion. We've got a panel reaction from industry experts and analysts, C.S. Carvella of ZK Research, Jake Callumbaugh, who's the managing director of cloud strategies, industry legend, been on all sides of the table. And of course, Dave Mulaney, investor, board member, former CEO of AVATrix, former CEO of Nacira, early employee at Palo Alto Networks, and Cisco, great pedigree and networking on the industry side now investing. And of course, Dave Vellante, co-host of theCUBE. Gentlemen, let's dive in and delve into this deal. Okay, HPE, we've been following their networking. All of us have been for years. HPE's had, remember, back in the old days they had switches and hubs and, you know, they've had a business. I wouldn't call it jumping out of off the page from a value perspective, but Juniper, world-class routing and switching, network connectivity, app development, service provider business, pivoted into software, putting those together, almost the same market cap. So let's get into it. Guys, what's the take on the deal? Zesh, you're in CES, we'll go to you while you have connectivity. What's your quick take on this deal and impact of the industry? You know, speaking of connectivity, actually CES doesn't offer pretty wifi. How weird is that? So I'll tether off my phone. You know, my take on the deal is, I understand the rationale, but if you look at both companies, the stocks have been going sideways, and we're both having a tough time growing. I do think between the two Juniper probably had some version of growth potential, you know, maybe low single digits coming up with their enterprises is now being bigger. But it's a tough industry, right? Networking is, for the last 30 years, it's been 1-800-pound gorilla, and then everybody else is much smaller. And so networking scale matters. And so if you're going to serve with the big global companies, Cisco's in the best position to do that because they're sized. So if you combine the two companies together, you get, in theory, a much bigger company that can compete with Cisco. With that being said, it just, I don't know, the timing of it seemed a little odd to me with the interest rates being so high and, you know, not that they paid for it. But I think, you know, overall, I understand the rationale behind it. I think there's a lot of complexities behind it. And you, I've never been a big fan of mergers or acquisitions for purely consolidating share because there's a lot of integration issues, channel issues, customer issues, you know, culture issues, and as you're doing that, the one plus one network comes out to three, right? It usually comes out to less than two. And so historically, when we've seen companies acquire for consolidating share, unforeseen problems that would come up. So this comes down to execution, we'll see. But I do think there's a lot of bugs along the way. Well, we'll know the mist and Aruba stuff will work when CES gets good connectivity and making it free. Steve Mulaney, you know, we talked, I talked about the deal with Dave, $40 a share, $14 billion, he thinks it's a creative. I want to put, we'll get back to the financials in a second, but, you know, this is the strategic implications behind the acquisition really focus on how it enhances HPE's edge-to-core strategy. What's your take? Because again, this is like the strategic aspects are consolidations happening, but, you know, bigger is better here, I guess, from if you look at some of the overlap, it's, you know, HPE gets some bigger, a bigger box player and service provider. They get maybe better switching and routing or, you know, where do you see that? I mean, Dave doesn't think it's diluted, but I mean, is it strategic in your mind? Yeah, I mean, I think AI is really, you know, we all know the infrastructure, you know, is dictated by what's going on in application perspective. And I think as AI starts implementing its way into every application, it's really taking the infrastructure from a very centralized architecture now in cloud and forcing a distributed architecture. And I don't know if HPE and Juniper see that, I assume they do. I know everybody's focusing on the missed, you know, portion of AI, but I actually think there's an even bigger story, which is, you know, the world is really moving towards AI and pushing it towards the edge, right? And we all know infrastructure is end-to-end system and networking is an end-to-end system. And I think we're moving from a very distributed, a centralized architecture to distributed, which is going to put more and more importance on guess what, routing, right? And so, multi-cloud, hybrid, edge decor, whatever you want to talk about it, people are going to need a multi-cloud, hybrid, distributed fabric. Of course, security is going to have to be built into this fabric, right? And it's going to be distributed, which is going to be high demands on that networking. And that's where I think the property of Juniper can really help. I don't know if HPE sees that. That's what I see. That's what I would do if I was at HPE. Sure, missed is fine. They didn't pay $14 billion for missed, right? Or at least they hope they didn't. And so, when you look at that, I think it's a big opportunity for them. Now, what Zeus just pointed out on a execution perspective, I lived or barely lived through the synoptics well-fleet merger. Merges of equals do not work, right? What do you think the first thing to Aruba guys and Juniper guys in networking, they're going to start fighting, right? That's just what humans do. And there's going to be overlap and there's going to be a bunch of bureaucrats that are going to be fighting for their territory. And that's going to happen for two to three years. So who's going to win in this? Just like who won out of synoptics and well-fleet merger, Cisco, Arista, Palo Alto Networks, Aviatrix, everybody else. So I commend them for trying to put strategically this narrative together. I think the challenge is going to be execution. And I think as always in these situations, it'll be the people that maybe are a little concerned about it right now, Arista, Cisco, Pan, but they will act. And I think they will be the ultimate benefit from this. Jake, you lived through the CEA Broadcom acquisition and you heard Zeus's comments. Are we in a different time now? And is that culture fit? I mean, it seems that Juniper and HP got a good culture fit. In my assessment, they kind of do have good cultures. I won't say super fast moving, but I mean very solid world class. Jake, you heard the comments from Steve on this cold battle. First, I think HP is not calling an emergency vehicle. I think that considering it HP buying Juniper, but to your point about Aruba and the fighting that could happen. Jake, what's your take on this and your assessment of the acquisition? Yeah, I agree with you that the cultures are probably pretty well aligned, but we've been talking in Silicon Valley a lot about the effect of post-zerp on startups. And I think this is a great acquisition that illustrates post-zerp on public companies, especially public companies that aren't gonna be able to create, listen, you can create cash flow for shareholders in two ways. One, you can grow the company through revenue growth. And the other is you can run the business more efficiently and distribute more cash back in the form of dividends or buybacks. Clearly, I think this is a consolidation play. It's the next crank in the turn of enterprise infrastructure consolidating, especially as more workloads, not only are they being built on the cloud, but they're also being shifted onto the cloud. We're starting to see cloud workloads go from what was kind of 20% to, I don't know if it stops 50, 60, 70, 80% of the enterprise footprint, but that necessarily means there's not a lot of oxygen left to support multiple large public company bureaucracies in the legacy platforms. And so, I think, I don't see where the growth story comes from here. When you look through both companies' businesses and their relative portfolios, it's kind of 0% to 10% growth. Over the next few years, they've both identified AI. And I think AI remains a very interesting speculative growth opportunities and wondering where that comes from on the infrastructure side. So, I think they're gonna have to drive a lot of their economics through synergies. And so, this feels much more like a Broadcom VMware play than it does somebody who's reaching for the future. Okay, let's unpack that. I think that's the key point. Okay, is it a consolidation or is there growth? Zs, Steve, Dave, Jake, what are you doing? I wanna pick up on something that Steve and Jake both said. I totally agree, Steve. There's a big test for Antonio Neary. I think he's up for it. This is a big nut for him to swallow, but I think he's got the experience for it. And I think, Jake, you're right on. This is a consolidation play. But Steve, I'm wondering, take us inside the board dynamics from both Juniper's perspective and HPE's perspective. If you had to make the case from Juniper's perspective, what do you think the conversation was like there and flip that, and what was it like inside the HPE board? Well, at Juniper, we're tired. We've had 10 years of no growth. We want out, right? And probably HPE's on the same time. We've gotta do something. We've been irrelevant forever. We're going to go into a relevancy. And there's this thing called AI that might be a lifeline that we can attach to that maybe can get us out of a relevancy. We've got to do something. Lord knows neither one of them can innovate, right? That they lost that 15 years ago, 20 years ago, right? They haven't innovated since Scott Crens. So, how long can you continue to go with no growth in the stock? Like at some point, you got to do something, right? And so I actually think this is not a bad thing. If I were there, I would turn it into a growth thing, but I think everyone is right. This is a consolidation play, right? I just think there's an opportunity for the growth play and I'd be curious does Antonio see a growth opportunity in this? Maybe, I don't know. Okay, so they justify the deal with a consolidation play. Dave and I just went through some of the numbers. It looks okay. The creative, if they squeezed and crank it, as Jake says, one more crank of the turn. Okay, growth, where's it come from? I mean, remember Juniper has had the cloud guys as customers on the big box side routing. You mentioned that, Steve, on one of their cores. They have routing. They said Contraia, which is an early multi-cloud play. Again, super early, maybe too early. They saw the pivot to software. Is this maybe the tail? They are the alphas in this deal with the internal Aruba? I mean, Aruba had great network in the edge, but the Scuttlebutt wasn't that strong. And there was a lot of, people were maybe just red at people complaining, but most people I talked to were like, some use cases they wasn't delivering. What are their gaps in Aruba? Is this a better pitch? Is there an upside with edge, silicon chips, and cloud? I mean, that's what I would only see as a narrative here. I think the most part, the Juniper products be the alpha. I actually think from a leadership perspective too, you mentioned Antonio, Dave, but I think Rami's the key here, right? So he's going to be, I actually think the structure of having all of network can roll into Rami is the right thing. I think this comes down to, when you do a merger, it comes down to how well you can execute and the reason they derailed to Steve's point is he went with a lot of infighting. Rami is a good execution person, right? I think he's, he took a couple of those largely telco, Augusta made an enterprise, Augusta squashed a lot of the people that complained in there and brought in new people and bought companies. And Juniper wasn't a very good acquirer pre-Rami or Raheem, right? And I think MIST worked pretty well for him. And I think 120 work pretty well for him. I think he has a lot of data center chops. And so to me, he's really the key leader in driving this to become a growth story versus one of consolidation and shedding products and things like that. But I do think from a product perspective, I know what you're saying, Steve, about the edge, but I do think MIST really was the crown jewel in this. And what I'm really interested in is can they take MIST though and make it a broader AI ops platform that could also help with a lot of the HP computing stuff. And then HP would then become, or MIST would become that control point for all things IT that fall into HP portfolio. They can do that. They might be able to use MIST as the tip of the arrow to be more of a platform play, which central never became that. Central was a rubric product for their networking. And I think one of the things MIST, they could use to be more of an HP product that's possible to live. So Jake, so to Milani's point about this consolidation versus growth, I mean, we've seen an innovation. We've seen the consolidation play work for growth companies like Palo Alto Networks, like CrowdStrike. Now they've got innovation as well as their tailwind. Do you think that HPE can pull off both either through as Steve saying, make AI a growth engine. They've got the high performance piece. Maybe there's an AI play there, but you can grow through with the consolidation play. That's the market trend, but you've got to have innovation. What's your take on that? Hey, listen, anything is possible. Whether or not it's probable is a different question. I think as Steve pointed out, the political war internally between these two organizations around who is going to drive a certain product category versus who becomes redundant and who winds up on the synergy list for this combined organization, that's not gonna be pretty. The real politic of how that plays out is not gonna be fun at all. I absolutely think there's room for innovation, but when you compare them to a player like Palo Alto, you know, Palo Alto is buying next generation bleeding edge capabilities as plugins. And the thing that Palo Alto did is broke the mold on integration, right? So the traditional story is that integration of emerging bleeding edge technologies into large platforms, doesn't go well because the people who own the budgets, the SVPs and EVPs who run these categories, they don't really understand how that business was built and then the innovators get fed up and look for greener pastures and wanna go play the startup equity game. So I think in theory, it's possible. I think the reality in how corporations are organized make it difficult. And Antonio's gonna have to do something different when he brings these companies together relative than what's traditionally done in large corporate technology bureaucracies when you have these kind of large acquisitions. The talent question's a good one, like brain drain and then infighting, right Steve? You brought the infighting up. The reason companies don't innovate as Jake pointed out is the talent leaves. I mean, Juniper hired a bunch of Google guys from where they had the contrail acquisition. Remember, they had some action going there. What does HP do to stay relevant? You said they can be, it's okay, but they're not relevant, but how do they become relevant? Does this give them a quarter edge? Is it better at the edge? What do you see if you're squinting at the deal from a strategic growth perspective? You asking me, John? Yeah. Oh, you and everybody. Yeah, I'd say it doesn't become growth. I agree with everybody what they're saying. The problem with what they're doing is it's very much focused still on the physical world of networking, right? Boxes, guess what? It's shifted from boxes to software and cloud about five years ago. And the problem is neither Juniper nor HP can spell cloud, right? They spell it CLWD, cloud, right? Somebody's got, you know, they won't have the growth until they actually start going after where the growth is, which is in the cloud. And you compare it to Palo Alto. When Nikesh came into Palo Alto, he pivoted the entire company overnight that said, we're all in cloud. Because before that, they were all on-prem boxes. And they thought, we're fine. We don't need to do this thing called cloud because they spelled it CLWD too as well. Nikesh turned that company around and now their $100 billion market cap with a lot of growth made a lot of acquisitions. I don't know if HPE sees that. I think they're still in the old world. So you won't see any growth. It'll all be kind of everything we've been talking about, just a consolidation play and we'll fire half the people across different organizations. And they'll be focused on that. The growth will come from when they really truly understand that this is a cloud-centric, cloud-first kind of world. If you see that, then maybe you see strategically, they get it. If you don't hear anything about that, then it's completely acknowledged. So compare that to, say, Cisco's growth strategy. Well, I don't know what Cisco's growth strategy is. Z, as you were covering both of them. You're by Splunk. Splunk strikes. I don't understand. The interesting thing about Cisco's acquisition of Splunk is that Cisco stepped forward into security and it also built on app dynamics. They are moving up a layer into application and security, right? And so I give them points for going into the growth direction. Still have to execute, still big companies, still have to work it out. The gaps in this deal, big questions I have is, HP's real strength is GreenLake in some sense. And so how does this improve GreenLake? I'm not sure it does, but maybe there are some things that can do with, when networking underneath that makes GreenLake more interesting with Juniper. And then- We agree on like bigger portfolio, right? Absolutely, I agree. I mean, if you GreenLake, now they got networking sewn up if they GreenLake it, but they didn't talk much about GreenLake. Because Wall Street doesn't care about GreenLake. Big platform right now. And we can argue about whether or not that's how much growth is there. But I think the other big gap or big question I have is security, right? So Cisco definitely is driving towards this cloud security, distributed security, security through data. So I think they get it again, have to execute. But I'm curious, I hear Juniper has a lot on the threat analytics side, but that's just one piece in a multi capability puzzle that they've got to bring together to really drive this networking security conversion story, which is powerful. Well, there are security assets out there that are right for the taking, but to Zeus's point, I mean, HPE's balance sheet is not that inspiring and that they're taking on more debt in this post-Zerp world. The interesting, isn't it, Jake? You got Splunk and Juniper, two big hosts of Zerp, M&A deals. But Steve, go ahead, sorry. No, I was going to say that, I think that the Cisco buying Splunk, everybody knew that was going to happen. They were just waiting for the right, they were waiting for Splunk to kind of get up off the floor. And so I think they did that. But the question is, we're in a new cloud world, as Steve says. And are there enough assets? Are they oriented enough to make that shift? And Steve, your point about you, if you were the CEO, you go after a growth story, because that's in your DNA, but that's a lot riskier for a company like HPE to take on. Think about it, what they go out and buy in Aviatrix, okay, great, they get innovation, but they don't get the revenue hit, right? They don't get the install base. Yeah. No, I mean, it's like Broadcom, right? Broadcom, from all of my friends at VMware, they have a very great strategy, which is they take zero market and product risk, they like version 8.0 of something, right? And I'm just gonna operationalize the hell out of it. And Hoc10, he's made a great business out of all that, right? So you're right. There's certain companies, they don't want anything to do with growth, they don't want any risk at all, right? I just want the plus one, version eight of something. And maybe HPE, that's what they're trying to do. They're buying revenue, that's what Cisco did was blunt, buy revenue, I get a multiple, I've got to show some growth and operationally, I'll make some improvements on it. Well, that's a great point about Hoc10. To your point, they have to do something. I mean, again, take us inside the board. What are the alternatives? I think that basically Antonio stepped back and the board stepped back and said, look, the opportunity is in networking. We're winning there with Aruba. Let's double down on networking. Where else are they gonna double down? I suppose security, I suppose data, other data management is sort of meh. So networking seems like a logical place to try to grow. One of the telling data points they gave in the industry analyst's call was that the combined company would be 31% of HPE revenue, but 56% operating problem, right? And so networking is a lot more profitable than compute. And I think we are moving into a network central world, obviously, right? And so again, I think we do is don't... Sorry. Transforming HPE from a compute company to a network company is good for HP long term. You can argue how they get there or how they should have done it. But this transition is something I think they need to do. What's your point, Steve? The else is... I was just gonna say, again, I don't know how much HP really knows this because they're not really into the cloud world, but if they're paying attention, the world's going from centralized to distributed and it's not distributed connectivity. It's distributed computing, meaning compute storage, networking, security with AI. This is not just VPNs to edge locations. This is applications. You're gonna run applications. You're gonna have computing infrastructure. So if you're HPE, you're like, you should be really excited because that means I'm gonna sell boxes out there and it's gonna be running applications and I'm gonna need networking and storage and security. Because of course, the security concerns as you now build a distributed infrastructure are even far greater. So the demands on this fabric, if you would, and it's not just networking, it's compute storage and networking, computing, which is when I was at VMware, that's what got VMware really elevated. When they bought Nacera, they could really now have the software-defined data center story, right? This could give HP really a story. The problem is they're nowhere in the cloud. So they're only at that edge. They now need to look at the cloud. Actually, Jennifer had Conrail. The end-to-end systems problem. It's not just point-to-point. I think your point about the cloud, they had a little bit of Conrail. They were doing multi-cloud. They were getting into that. It's an open-stack action going on. If you remember back, and Randy Baez used to work there. We interviewed him on theCUBE. But I want to bring up the point. You brought up ZS around Margin. Dave brought that up as well. And at Reinvent, you saw Nitro. Networking was so integral in the new architecture. Steve, you're bringing this up on the multi-cloud scene as well when you guys doing a DVA tricks. And then the Hop-Tan connection kind of brings up another point. Networking is very sticky. So if you can operationalize the HPE and networking, there's not a lot of switching costs here. So it's a very Broadcom-like vibe. Broadcom, they're all about consolidating, make it easier, bundle everything. You want vSphere? No, you can't just buy that alone. You got to buy VCF, which is everything, right? Which includes NSX, right? So to your point, I think networking is prime for stickiness and margin expansion. By the way, his first thing that he looks at is stickiness. Exactly. Who are in the room with him. The other thing about Hop-Tan's model that'll be interesting for Antonio is that he was the guy to make all the calls on what groups and products were going to stick and who would lead what. So he was a one-man synergy machine. And I know a lot of people who stood in front of him who said he's very, very good at it. Like he's just, he's perceptually able to figure out who's got the chops and who doesn't and how to operationalize it, as Steve said. So it'll be interesting to see how Antonio handles that tactical component of the integration in a committee-based company. Well, let's see, to the former point, Jake, does the question to the panelists, does Juniper have that type of stickiness in that lock-in or is it really just sort of a alternative to Cisco? I mean... No, I think Juniper's got pretty good customer loyalty. I mean, their customer base is a lot smaller than Cisco's, obviously. But through MIS, they have picked up a number of really large enterprise customers. They won Walmart and Costco and some other big ones. So they do have, switching network vendors isn't an easy thing to do. No, it's not. And so I think their stickiness actually is pretty good. They've always had good, you know, a good relationship with the cloud vendors and service fighters and those markets are in the soft right now. But the timing of this, you know, is, I think, somewhat favorable, I guess, financially in that the whole network sector is in a bit of a bare gap right now. You know, Cisco warned that a lot of the customers that are just digesting products are just pretty, you know, pre-announced as well. You know, Juniper saw, they said they saw softness. So if HP was going to do this, I guess right now would be a good time because all the stocks are depressed, where if they waited, you know, maybe a year from now and networking came with a little bit of a rub-tick, then, you know, perhaps the premium would have been a lot harder. Well, let's unpack the switching costs real quick. Steve and Ziaz and Jake, what does it take to switch from Juniper to Cisco? I mean, Steve, you highlighted earlier about your days at Synopsys and Wealthly that became Bay Networks. I mean, you remember, it's not like you don't unplug a router and then bring in another one. I mean, it's the switching cost bar is high. What is that bar that people costs? Nobody's switching and I think as Ziaz said, people that are picking Juniper like Juniper, right? Their definition, by definition of the ABC, right? Anything but Cisco, right? So they're never going to, Juniper hasn't been growing now, but I think the customers that are Juniper customers like Juniper, Juniper's always had great product, right? So I can't see anybody changing. So the consolidation works, they got some stickiness, networking's the core of the edge, this is good for HP from the numbers, debt can finance it, consolidation play, optionality for innovation yet to be seen, but potentially there. Well, Ziaz's point about the debt loads, I think comes into play around how they find growth in innovation because the amount of money they have to commit to interest payments and an entire interest rate world, that comes directly out of BU budgets, right? And so the ability to find and allocate innovation dollars smartly is also going to be a key capability that Antonio's gonna have to figure out as part of this deal. I think if there's one bright spot about this and is that I definitely see a strong automation and consolidation through AI story on both sides from Juniper and HP, and if they can bring that together, the continued automation of complex hybrid enterprise architectures, that is a plus. And if they can rationalize the different, like I looked through their acquisitions over the last few years and they bought a lot of interesting stuff, the midst and some other things on both sides. So I do think if they can figure those product categories out, create some connectivity across the product lines that matters to customers. It's not sexy, but I think driving tactical success at that level is one of the ways that they can find success in this deal. Gentlemen, thank you for this emergency pod. Cube pod, Cube broadcast, appreciate all your time. Thanks for weighing in. Final question, let's end this with a lightning round prediction for 2024. Zias, we'll start with you. What's going to happen in networking in 2024 and what's going to happen with this deal? Well, they said the deal wouldn't close until the very end of 2024, 2025 or so. I'm guessing it'll get closed some time. But I'm guessing later in the year than earlier at the year. For networking though, what the vendors have said is true. A lot of customers bought a lot of product or put orders in when there were supply issues. They've now taken all their product in and they do have to consume it, figure out how to deploy it. And I think that this air gap we're seeing right now is probably going to last, my guess is, at least through mid-year, before we start seeing some growth back in networking. We do have another Wi-Fi sector coming up. One thing about networking is the speeds have gotten so high that I think it has elongated networks and network refresh cycles. So that's something the vendors got to work out. Zias, explain air gapping real quick of what you mean by that term. Well, the customers have pre-bought a lot of product. When there was supply chain issues, the customers got into a bit of panic. People coming back to the office, they need to refresh their networks. And so they took in a lot of product and that was reflected in most of the vendor numbers last year, they all had years. Now the customers have this product and they have it deployed again. And so they got to figure out how to consume it, how to deploy it, how to use it. And so now this growth that the vendor has seen that a lot of customers would have paused on things until they can get this, you know, one of the new stuff up and running it. And that wasn't just Cisco saying that. I mean, all the network vendors have said that. Jake, your predictions for the 2020, but what does this do to the industry landscape? You know, the deal's going to close into the air. Obviously it's going to flip a bit in the industry. You know, the consolidation plays, you saw Broadcom, CA, you've advised, you've invested. What's your take on networking? Yeah, I think, you know, we'll see more of this. You know, I have to think about kind of the remaining players, but you know, this deal feels like it sits in between Cisco Splunk and Broadcom VMware. And so, you know, I think the bigger getting bigger and it becomes about operational execution. And I think to make this deal successful, I think Steve's right, you know, they have to figure out how to incorporate cloud and really make the hybrid cloud story a reality versus just being kind of a hardware layer that, you know, could ignore cloud up until now. Dave, what's your take? Well, I think from a customer standpoint, customers want to simplify their vendor list. If for no reason that they want to, actually Steve was saying they want to do some innovation. They want to, you know, tap some of the new wardrobe and prune some of the old wardrobe. And so, I think networking, the prediction is networking and security are going to collide this year in a bigger way. And I do think the rich get richer. You're going to see some pickups and some M&A in security and networking. And I think that's probably a good thing for customers so that they can innovate on some new stuff. Change that wardrobe. Never fight fashion, Dave, right? You don't want to wear the 70s outfit in the 80s. It's punk rock, it's punk rock. Steve, what's your take on the deal and the impact of the industry? My prediction is this will be the year of cloud networking and no surprise there. I think HPE and Juniper are not dumb. They're going to figure out that this is the piece they're missing. I think Cisco's already figured that out. I think Arista even will figure that out. That's, you know, it's no longer about boxes in the campus at the Edge or in the data center. That there's no growth there, right? So you're talking single digit growth or you're fighting someone to do market share replacement which is a bloody battle. No one really wins because it's really hard. No one ever rips and replaces their network. It's just not done. So I think they're going to figure that out. And then I would also say, don't count Palo Alto. Networks is going to truly become what their name says. Palo Alto Networks is not a networking company. They're going to need to become a networking company because network security must be integrated into the fabric of the network and not a bolt-on like it was in the last generation. When you're in the cloud, it's not a separate thing. It's all together. It's got to be integrated in. So I think it's going to be a battle royale. You know, you've got my old company, A.B.A.Trix leading this, right? And it's a great opportunity for them as well as some of the incumbents. And I think that's going to be the battle this year. I have to get one. Sounds like you're predicting a Palo Alto Arista title. Well, that's why I got one more question. You guys just prompted another one in my head. You got a prompt engineering thing going on in my head. Okay, quick, final question, final, final question. Impact to the board rooms of those other guys. Cisco, you mentioned Arista, you mentioned Palo Alto. They see this news. What's the huddle? What's the slides look like? What's that conversation that's boiling over right now at Cisco, Arista, Palo Alto, or competitor? I think for competitors, it's up on the gas. Whenever, if you look at some of the comments, Reddit, even there was an Aruba customer on my LinkedIn post that said, we've got some concerns, is look, if there's uncertainty, that gives the competitors the opportunity to jump in there and try and, you know, feed off that uncertainty. And so if I'm in the board with the competitive companies, I'm jumping all over this and trying to feed the fun that's likely, you know, being created or at least the concern the customers have. And by the way, we talked about Palo Alto. I think Fortinet has done a really good job of networking to their portfolio as well. They're one of the big SDUWAN vendors, and I just don't want them to go unnoticed because I think they've got a pretty big stake in this conversion of their security as well. Steve. I think Cisco's got its hands full. They got a lot going on. So I don't think they're going to get too distracted by this. I think Arista is interesting. I think, you know, Jay Sri's got great growth engine going on, but, you know, at some point, they're going to have to continue on TAM expansion. And that brings up some opportunities that we talked about in M&A, but, you know, they can be selective right now. Well, they're small enterprise, right Dave? So that's where their big growth engines is becoming more of an enterprise player. Absolutely. They don't jump into things without thinking it through. And they're really a well-run company. What's the board like? Steve, Jake, take a stab at that at all? Yeah, I would just say, you know, you've just mentioned Pan and Arista, you know, and Jay Sri's done a fantastic job, but, you know, her market cap's 75 billion. Let's not forget. It's not 15 billion anymore. It's huge, right? 75. It's huge. Are you going to go to 150, right? And if you're Pan, how are you going to go to 200 billion? You're $100 billion market cap right now. How are you going to go 200? You ain't going 200 by getting another 10% of the security industry, right? So they're going to have to make big moves and they're going to have to go after, you said TAM expansion, John, right? They're going to have to go after industries. And, you know, both of those people are going to be looking probably at HPE and saying, you know, maybe we ought to be doing something like that. We're going to have to go after the entire infrastructure market and not just either networking or security. We're going to have to do probably all of it, you know? And by the way, compete with the cloud vendors, right? So the cloud platforms, you know, Steve, you brought it up. The networking in the cloud is becoming a big deal and they're starting to create their own, you know, enterprise cloud platform, cloud networking platforms. So how to create relationships and how to create compelling value for customers relative to those cloud guys and make sure that they can stand on equal footing, I think is an imperative for all the traditional infrastructure players. Gentlemen, a great conversation, exceptional analysis. The best on the industry I've seen out there. Clearly you guys are awesome and thanks for being part of our CUBE community. Again, real time news going down. We appreciate you taking the time to pop into the CUBE. Emergency CUBE, appreciate it. Thanks, John. All right, guys, thanks so much. Again, breaking news here. We're on it with the analysis from the CUBE Research and its community, the CUBE Collective. I'm John Furrier with Dave Vellante and guests on the panel. Thanks for watching and keep in touch.