VMware Explore 2024 | theCUBE Pod Analysis

Clips
News
More from VMware Explore 2024

Dave Vellante

Co-Founder & Co-CEO

SiliconANGLE Media, Inc.

The CUBE Pod analysis: Navigating VMware ecosystem challenges and strategic shifts at VMware Explore

There are several interesting power dynamics currently at play in the industry, particularly around VMware ecosystem challenges. This year’s VMware Explore event has been marked by the intersection of emerging technologies, such as generative AI and shifting strategies among industry giants.This year’s VMware Explore looks very different compared to years past. It’s sparsely populated, but it represents the tip of the pyramid, according to Dave Vellante (pictured, right), chief analyst at theCUBE Research.John Furrier and Dave Vellante discuss power dynamics during a CUBE Pod analysis at VMware Explore.“The customers that I’ve talked to here are all investing in VCF

play_circle_outline CEO Transition at AWS and Impact of NVIDIA and CrowdStrike Earnings in Market Dynamics

play_circle_outline Analysis of VMware Explore 2024 and market news

play_circle_outline Reflection on VMware's ecosystem and partnerships

play_circle_outline Evolving role of Q in Amazon's cloud offerings

play_circle_outline Importance of scalable applications and what's next after SaaS

play_circle_outline Navigating Amazon's Cloud & SaaS: Growth Strategies, Data Management, & Adoption Considerations in the Cloud Market.

play_circle_outline Focus on vertical industries

play_circle_outline Amazon's cloud strategy for on-prem

play_circle_outline The power-law of AI and its implications

play_circle_outline Revolutionizing Supercomputing: Democratization and Kernel-Level Programming by Software Developers

play_circle_outline NVIDIA's strategic position in the market

play_circle_outline Competition in the silicon and hardware market

play_circle_outline Bubble bursting in tech industry

Info
Transcript

theCUBE Pod Analysis

Dave Vellante

Co-Founder & Co-CEO SiliconANGLE Media, Inc.

HOST

The VMware Explore 2024 coverage of theCUBE with John Furrier and Dave Vellante focuses on market news analysis, including discussions on AWS CEO transition, NVIDIA and CrowdStrike earnings, IPO markets, semiconductors, and the Fed. The vibe at VMware Explore is quieter this year, with a focus on VCF and critical applications. Opportunities for partners to collaborate with VMware are highlighted due to the lack of ecosystem presence. Analysis delves into cloud market growth, vertical industry solutions, and Gen AI implementation on-prem. Discussions also touc... Read more

explore Keep Exploring

What topics were discussed by Matt Garman during the conversation? add

What is the vibe like at Explore compared to previous years? add

What companies typically have a larger presence at this event and why might some be missing this year? add

What is the current innovation opportunity and the impact of the Gen AI moment on infrastructure work, particularly in relation to Broadcom taking over VMware two years ago and the importance of moving to AI with security and cloud operations in enterprise customer strategic visions? add

What are some potential areas where SaaS apps are being used to solve more complex problems, and what role do supercomputing and the relationship between cloud and on-premise solutions play in this trend? add

What is the current market growth and strategy for AWS in relation to the adoption of cloud technology, especially considering the relatively small percentage of IT spend currently in the cloud? add

What is the cloud strategy generally, and AWS's strategy specifically, for going on-prem? add

What is Amazon's strategy for the data center market and their approach to cloud operations and distributed computing? add

What is the power-law of AI and how is it impacting the digitization of our world? add

What did Garman agree with regarding the democratization of supercomputing and the development of ecosystems on top of hardware? add

What is the significance of NVIDIA millionaires in Silicon Valley and how has the stock performed since 2019? add

What company has a killer business model in the silicon and hardware industry by focusing on connect centricity and has locked up the market for the best silicon for interconnectivity? add

What potential buying opportunities may arise when the bubble bursts, particularly in profitable companies like cloud companies, NVIDIA and Broadcom? add

bolt Powered by CUBE AI

theCUBE Pod Analysis

search

>> Welcome back everyone to the VMware Explore 2024 coverage of theCUBE here. I'm John Furrier, your host of theCUBE with Dave Vellante, also host of theCUBE Podcast, also known as theCUBE Pod, and we are going to spend the next hour doing a CUBE Pod session analysis of the market news today, what's going on. Dave, great to see you.

Dave Vellante

>> Is this our CUBE Pod for this week? Well, let's do it.

>> We'll do this together. But I think this is important that we bring theCUBE Pod kind of vibe to this stage we're on. Obviously celebrity appearance with theCUBE Pod here at VMware Explore, but there's so much to talk about, the vibe here at VMware Explore and the industry. We had our opening session to kick it off with Rob Strechay from theCUBE Research. There is so many power dynamics going on in the industry that we'll talk about that go well beyond Explore. So this segment's really to look at what's going on in the news today, this past week, what's coming up. I just posted this, 'cause with Matt Garman of the new CEO of AWS, the Bill Belichick, I called him, to the Andy Jassy, Bill Parcells. Can he go on a dynasty building run?

Dave Vellante

>> Check that out on SiliconANGLE.com, it's on your Substack, it's on your LinkedIn. It's a must read. Getting inside baseball on how Amazon is going to evolve over the next five years.

>> And Matt Garman really opened up and shared a lot, there's a lot in there. The video's going to be... I'll do another story, follow on. There's a whole nother, I think about developers, how application is going to scale. Scalable applications, what's after SaaS? We're going to talk about that here actually too. NVIDIA, as we discussed, NVIDIA earnings. CrowdStrike earnings coming up. The IPO markets are hot, private markets are hot, semis are hot, and then obviously the Fed was at Jackson Hole last week. Rate impacts. Also, crypto. I like the crypto side hustle there. Fancy football day, while the pools are hitting, Snowflake, LinkedIn .

Dave Vellante

>> You in this year? Are you in?

>> I got in three pools. I got in theCUBE pool this year. Christian hooked me up. So I missed it last year. That's fantasy football, not a Pick'em league. I mean there are two Pick'em Leagues and one fantasy. And then finally, Dave, your AI agentic post blew minds this past weekend. I want to unpack that because there's so much in there to read between the lines. And of course Wall Street Journal, CNBC talking about data constraints and M&A. We're going to unpack all of this on this segment, so you really want to connect the dots here, theCUBE Pod is going to break it down. So let's get started, Dave, the vibe here at Explore.

Dave Vellante

>> Yeah, it's a completely different Explore like one we've never seen before. It's always usually, as you know, about the ecosystem and it's huge 15, 20, I think at some years they were claiming 25,000, Moscone, obviously they did Vegas as well, but this is not packed. I mean you can see it's pretty sparse, but it's the tip of the pyramid. The customers that I've talked to here are all sort of investing in VCF. They're sort of what I would call the all-in or mostly-in customers that basically are running mission-critical applications on VMware. And so the vibe is completely different. It's much quieter. I kind of like that in a way, it's sort of relaxed. You know what it reminds me of John? Remember the SAP Sapphires, it's sort of like a smaller version of that, very quiet, sort of stayed business, not like big loud keynotes and bongo drums and people dancing and spend a lot of money on the glam. It's just, let's get down to business.

>> Remember 2012, I lost my voice at VMware. They put theCUBE stage right at the edge, high up on a stage where all the sound from the expo floor came, and we were like scratching our throats. I stayed out one night until 4:00 in the morning to get the scoop on the 3PAR, HPE acquisition.

Dave Vellante

>> Remember that?

>> Finally got it out of the guy to reveal the source, and then we broke that Frank Quattrone story where he had the 3PAR deal.

Dave Vellante

>> He had the auction going with Dell and HP.

>> It was fun times, but that was over 12 years ago. Fun times. Now it's changed. I think you're right. This is a business and community show, because there's so much risk-reward opportunity on the table of VMware right now because VMware runs in a majority of the enterprises out there, IT departments. And there's so much built around it, jobs, people make money from it. There's software developers, there's ecosystems. So whether you're just a person who has a job, if that goes away, you're out of a job, right? So there's so much at stake and it's the most watched piece of this show this year as what will happen with the ecosystem. Are people going to switch? What's the deal with the licenses? Is VCF ready? Is it truly a great platform? Can it sustain another run of durable value creation in the IT world? I'm telling you right now, that's all everyone's talking about.

Dave Vellante

>> Well, it's interesting. I mean there are obviously sponsors here. I mean, an example would be Red Hat, but the ecosystem is not here in force. Normally you would see Dell, HPE, NetApp has somewhat of a bigger presence here, Pure, et cetera. They're not here in force like they usually are. Infinidat actually is doing some stuff tonight. I saw they've got some decent customers inside the VMware base. I actually think it's perhaps a missed opportunity, John. I think some of it is this emotional backlash, certainly from partners that didn't make the cut. Obviously they're not making the cut so they're not partners. But the ones that did make the cut, there's real opportunities here. And here's the opportunity. You know how Jassy, not Jassy, Bezos has attributed the quote, "Your margin is my opportunity." Jassy told you one time in your private interview with Andy, "I'm not sure that Jeff ever said that." But anyway, the industry gave him that because that's what Amazon has done. So like blockchain, well crypto, we'll attack the middle man. But anyway, to me it's an opportunity because Broadcom is setting a high umbrella on margins and pricing, and most of these services companies, these partners, they're operating on much thinner margins. I mean even a company like Dell, it's operating margins is in single digits. And so Nutanix, take Nutanix, they're operating margins are virtually nonexistent and so it's an opportunity for them to actually partner, however that partnership evolves, with Broadcom to pick up the slack that Broadcom is less interested in. Broadcom wants the customers that are here that are all in that are durable. Now of course they've got some options in other SKUs, but if you're just sort of not a huge VMware customer or you're not all in and you're not profitable to Broadcom by the way they measure profits, exceedingly profitable company, well then the world's your oyster because you can pull away, compete for I should say, some of those workloads. And you should compete for them, whether it's Nutanix or Red Hat or SUSE or HPE, Ubuntu, some of the open source stuff. It's all there for the taking. Take it. Why not?

>> Yeah, I mean I think the innovation opportunity right now is critical and I think the Gen AI moment is causing so much infrastructure work that two years ago when Broadcom took over VMware, I'm sure they saw it on the chip side, but I don't think they realized the impact of it. Maybe they did, but it's going to be a factor in risk management for how they roll out and get that platform durable for the next generation. Strategic vision of the enterprise customers right now that we talk to and it's coming out of our research is if you're not moving to AI with massive security built in with full cloud operations, you're nothing. That's why I think VCF hangs together, and in my conversation with Matt Garman and the exclusive I posted this morning, you'll see him open up and he really say, "Look it." And I compared him to Bill Belichick and Andy Jassy-

Dave Vellante

>> I love that....

>> because Matt Garman was one of the original product managers for AWS. He worked hand by hand with Andy. Ironically, was at a hockey game with both Andy and Matt Garman, and I'm like, Andy's bringing Matt Garman, that's weird. He's a sales marketing guy, okay, whatever. So I've known Matt for years, right? So-

Dave Vellante

>> He was running the defense in a way.

>> So EC2 is, I compared Belichick was the defensive genius, as was Bill Parcells who loves smash mouth football. And you know the history with the New York Giants. Parcells was the man. He had great lineage of great coaches that came out of his organization, Holmgren, a bunch of other great coaches. So Jassy could be that lineage leader. I think he might've lost a few people. He's got to rein them, boomerang them back in. But Garman was his right-hand man, did EC2, their core service, that's where all their money comes from. They've been grooming him, but now Selipsky's out, he's out, he could land somewhere else.

Dave Vellante

>> But he could have been Pete Carroll. You didn't say that out of respect, I presume, but he could have been Pete Carroll, right?

>> I didn't want to muck up the story, but I think-

Dave Vellante

>> Yeah, that would've confused things....

>> he's Pete Carroll.

Dave Vellante

>> But think about it. I mean, Pete Carroll, great coach. He had to follow the great Tuna, the big Tuna, and just wasn't going to happen. But then he went on to be tremendous at USC and won the Super Bowl with Seattle.

>> I tell you, I was a huge fan of Adam Selipsky. When he came on, I've had multiple meetings with him, we've done exclusives. The guy's great. He's a great executive. He's no chump. He's awesome. The timing of the market shifting and how he got brought in with the Jassy leaving was just really weird. To your point about trying to follow Bill Parcells, Carroll came in when the big Tuna, Parcells, had the blowout with Robert Kraft. So Pete Carroll was not in a position to be successful. I kind of see similarities with that with Selipsky, because he could have been successful, but the way once it started shifting, it became... And then all the narrative around Gen AI, that's why in my story I put the clip onto Cincinnati because-

Dave Vellante

>> That's beautiful....

>> I think Garman is kind of pulling the Belichick by saying, "Are you Gen AI? Onto Cincinnati." His version of that is just customer, customer, customer. So he's kind of Belichick-like in his tone, all the reports. "Are you sure you're getting Gen AI right?" All they want to ask about is how's Tom Brady doing? How's Tom Brady doing? Is he going to come back? Is he going to be suspended? So Tom Brady is Gen AI, offense. The defense is the infrastructure. So Microsoft puts a good offense out there with OpenAI, but their defense sucks because their infrastructure's not strong. And so I think Amazon has a strong defense. So again, it's a weird NFL analogy if you don't follow football. If you do, you'll get it. If you don't, you can look it up. But I think you and I talked on our CUBE Pod last time, Amazon has an opportunity to balance the game and they got to get offense. They have to find that Tom Brady moment. Who on their organization, what application, who becomes that unsung hero? And Brady was not on the radar. He had the work to get in, but once he got in he was good. So I think that Gen AI is the offense for these clouds, and the defense is you got to maintain the security you need scalable systems, clustered systems, custom silicon. Amazon crushes that, they are great at that, and now they're building up the Gen AI, Dave. So to me, this is why I wrote the post that way.

Dave Vellante

>> John, 10 or 12 years ago, you and I sat down with Jerry Chen at a re:Invent and asked him back then, "Do you think Amazon will move up the stack and start to develop applications to compete in that realm with Microsoft?" And he said, "I don't think so. I think basically their strategy is to provide the infrastructure." I mean, Amazon's essentially a hardware company, if you will, but with highly automated services around it. And he said, "I think they're going to enable their ecosystem to build those applications." What's your take on Q? Did that come up in your conversation at all? Because it's a huge opportunity for Amazon. They don't really compete in SaaS, I mean it's very tiny.

>> No, he did bring it up, it's in the interview, but I didn't want to go drill down on on the story because Jassy posted-

Dave Vellante

>> And it detracts from-

>> Well, Jassy just posted the day before, if that story's out there, that developer's going to start building.

Dave Vellante

>> So what's your take on Q as the potential to be Tom Brady?

>> Well, I think the combination of Q and humans to build these scalable apps, I think the next Tom Brady has put some points on the board. Now remember the defense, remember the Patriots, they had a good defense, but Belichick had to scheme the defense with a salary cap, and he put all his eggs in the offense. So the Patriots are known for offense. So to beat the Patriots, you had to score more than 45 points, 'cause they were going to put up 45 points.

Dave Vellante

>> Charlie Weis, Josh McDaniel.

>> So I think Amazon has to use Q on any lever they can to create offense, and to me it's going to be about the new applications. I asked Garman that and I said, "What's after SaaS?" And he said, "Scalable apps." And when you look at the hard problems, it's not website, web application to SaaS on the cloud, that's what Gen-1 was. Now it's SaaS apps to solving more harder problems. So if you look at the supercomputing being democratized like with NVIDIA and AWS, the next killer app is major biology, computation, major problem solving. So I think you're going to start to see the killer apps coming around quantum and super cloud kind of capabilities. And again, all that doesn't happen until the theme we've been discussing about the relationship between cloud and on-premise gets resolved. And I think if you look at all the action right now, even though Amazon doesn't want to talk about on-premise, it's still cloud operations. So the relationship between cloud infrastructure and on-premise is huge context to the Gen AI, AI movement, because all the data's on site and or in the cloud, this has to be resolved very rapidly. And again, all the work is being done on the data center side and that's going to be where the action is with GPUs. But GPUs are changing, so you got... Again, what I did like about the Garman post, Dave, is that he did give me something new. And that is that he laid down his inference opinion. He believes that inference is huge. And to quote him on that, he says, "Why would someone do all this training and not have inference?"

Dave Vellante

>> Yeah, they go hand in hand. I mean, Jensen said the same thing.

>> Inference is in the application, of course we called it on theCUBE. But look, I like Matt Garman, I think he can be that Belichick. Right now he's the defensive coordinator taking over the head coach job of the new organization. I think he's got to bring the offense, he's got to find his Tom Brady moment, and that's Gen AI. It's very clear to me that that's the offense and then the infrastructure's the defense, and that if they can do that, then they'll continue.

Dave Vellante

>> So the reason I bring this up is because as we all know, Amazon started the AI, the IS, the modern IS and cloud trend, and of course Google and Microsoft jumped on board and have competed. Amazon does virtually no SaaS business other than it's got a little bit of Chime and Q, does some stuff through its partners, certainly through the marketplace it's got a good software presence. But we've seen, and we've called it and we've reported on, written about, researched and called into question some of Azure's numbers. I mean everybody just takes, Charles Fitz just basically takes the numbers that Microsoft claims for cloud and says, okay, that's Microsoft's cloud business. Everybody does the same thing for Google and then they take AWS's business. Well, AWS's businesses, it's all the infrastructure. I mean, it's almost like comparing Dell's business to someone like Oracle that has a huge software business. It's not apples to oranges. So we've tried to strip out all that SaaS and we've done some work to do that. Microsoft, I presume you saw this, revamped its reporting structure to give better visibility on cloud consumption revenue. Okay, so Microsoft had always included per user revenue in its Azure reporting for things like mobility and certain security capabilities. They're changing that, they're stripping that out and then they're bringing in, I think, advertising, which is kind of apples to apples with AWS. I don't know, who does more advertising? That was kind of a weird thing. But also the cloud consumption stuff that they're doing. So that's really an interesting change. The reason I bring that up is we're going to start to get a better picture of what an apples to apples cloud market looks like. And right now Amazon has virtually 0% market share in SaaS, so it's an opportunity for them. It's not a threat in any way. It's really an opportunity. And so to the extent, I'm really interested in whether or not, I don't want to call it a change in strategy, it's an evolution of the strategy. It's a line extension, as you sometimes like to say, into SaaS, and that is a multi tens of billions of dollars of opportunity for Amazon to do things in vertical. They've started to do some of that with Connect. And can Q be that platform on which either Amazon and/or its partners build these scalable apps? That could be the Tom Brady.

>> Yeah, I think that's a great analysis, but I'd also say that that's one element of it. They've got the horizontally scalable data layers. What happens with the relationship between data on-premise and in the cloud, has to be cloud operations. So again, I brought up this up in our session with Rob Strechay, what's the cloud market growth and strategy for AWS, for the adoption to continue, especially when only a small percentage of the IT spend is currently in the cloud? So there's still more growth opportunities that-

Dave Vellante

>> I don't know about that.

>> What percentage of IT spends in the cloud right now, would you guesstimate?

Dave Vellante

>> Well, I would say that 30 to 40% of the workloads are in the cloud.

>> What percentage of spend?

Dave Vellante

>> So most of the spending-

>> Just guess, throw a dart at the board.

Dave Vellante

>> I think it's more than the 10% that Andy Jassy says. I'd say it's closer to 30%. I think that's a big number.

>> That's pretty closer to workloads.

Dave Vellante

>> Okay, so it's not 50. It's getting there.

>> Well, exactly.

Dave Vellante

>> But remember too, most of the spending in IT is in services. That's at least half the spending. So if you say IT services is 10 billion, take out half, which is professional services. Now over time with AI, you can go after those, that professional services market. No question about it.

>> If you back out services, I think just raw horsepower, that's a smaller percentage, I think there's huge growth in the cloud still. I'm a big bull. I'm long on bull. All clouds.

Dave Vellante

>> Well, no, there's no question it's... Well, look at the numbers. The big three cloud vendors, big four if you want to throw in Alibaba, is growing at 20% a year. On-prem's growing at single digits. Look at Cisco, look at Dell, look at IBM, look at HPE.

>> My point though, well first of all, I'm not disagreeing, I agree, but there's so much growth. Then to your point about vertical industries, I think that the industry-specific solutions that are going to get more Gen AI is in critical. Some of the public companies that have data, I just saw example where ICE, the parent company of NYSE just randomly noticed that their earnings are up big time because of Gen AI on the energy, these regulated sectors. So I think this is going to be kind of like a, I won't say a big story, but I think it's a big story where there's going to be this growth in these other markets. It's not just general purpose. So I think Gen AI is going to have great value in these markets where the retrofitting of pre-existing workloads are going to get immediate value. I think that's where the action is there. So love this vertical market specialization. I think that's going to allow them to set the table for transformational and innovation, because everyone's looking for the shiny new thing. What's the big bang result of this is the new thing, like ChatGPT? I think it's more of retrofitting existing.

Dave Vellante

>> But let me ask you, do you believe in the Gen AI power-law that theCUBE Research put out-

>> Yes....

Dave Vellante

>> in the last year that you obviously helped develop? Of course you believe it.

>> Of course I believe it, I helped develop it.

Dave Vellante

>> Isn't the corollary to that, that a lot of that is going to occur on-prem?

>> I think in the long tail and in the torso it's going to be on-prem and I think as you move up the power-law, it's going to be in heterogeneous environments, because look it, let's just be clear-

Dave Vellante

>> Wait, wait, wait. Stick with this, before you go there. If that's the case, and then I'll leave an open question and you can maybe answer it in your response, if that's the case, then it comes down to what is the cloud strategy generally, and AWS's strategy specifically, for going on-prem? I think Microsoft with Arc and Azure Stack, the old Azure Stack, has a very clear strategy there. It's not as clear to me, other than Outpost, what Amazon has there, but please make your point.

>> All right, so I'll go back and say why I think what the clear stretch of Amazon is to our mainframe conversation. I think the clouds are the mainframe and they're open heterogeneous mainframes. I think if I'm Amazon, I want the data center market to become cloud operations and hybrid and distributed, but ultimately running all the stuff into the power source, 'cause they're going to have the GPUs and the silicon, the defense as we say with Belichick, and I think that's their strategy. Run through their VPCs and also allow data processing to go wherever it is. And if the customer wants to do data processing on-premise for privacy reasons or governance reasons or whatever reasons, Amazon will let that happen. I think it's better to have it develop organically than jam Outpost down everyone's throat. Instead, Amazon will shift and they are shifted to local zones, getting better infrastructure. Amazon's core competency is infrastructure as a service, so they're clearly investing all that in there. And then making sure they have a robust higher-level services platform as a service to make it all work. That's the databases, the data pipe-lining and they got work to do there. They're not perfect.

Dave Vellante

>> Are co-los a threat to that? In other words, if I could take an Equinix, load up an Equinix data center with my data and then I can be proximate to an Amazon cloud-

>> It's kind of like saying a Snowflake competing against Amazon. It's going to drive more services. Again, if you're playing the long game with Jassy and Bezos' mindset-

Dave Vellante

>> I say go for it. Yeah....

>> I'd say let that develop, because we're going to get that business. Their job is to get the business back. So I'd say in the short term, myopic view would be, oh my God, on-premises is a problem. Amazon's sticking to, onto Cincinnati, Dave.

Dave Vellante

>> Okay, so now the other piece of it that I want to ask you-

>> Hold on one more point about the power-law, because I think the power-law... If you don't know what we're talking about, search the power-law of AI, it's on SiliconANGLE and theCUBE Research. We basically put out a premise over a year and a half ago that there'd be a power-law of AI models. We were right, the specialty models are in the torso, and then proprietary models or IP-based data sets are in the tail, on-premise for companies. That's what everyone's talking about today. Why I'm bringing this up is let's be clear, the digitization of our world is happening and you're looking at neural networks coming, you're seeing new kinds of retrieval. Agentic, what you wrote a post on and we've been talking about here in theCUBE and I've been talking about it from a clustered system standpoint is as things get digitized into data, you now have a whole nother world. That's why I'm big on blockchain and decentralized infrastructure, because ultimately whether it's Bitcoin currency or not or cloud services, everything will be digitized and digital. That means the power will be shift, you'll see commoditization of stuff and the rise of new economics like agentic systems and new kinds of data models. So we're going down that road. So the power-law will continue to make the play and that's why I think Amazon will be at the top of the power-law with the big clouds, including Meta and others. And then as you go down, you'll have heterogeneous environments and roles like the data center or Equinix or the hosting providers or the CoreWeaves of the world. So you're going to start to see all that. There'll be a lot of bubble bursting, but I see the transformation clearly. But at the end of the day, people are just trying to make IT work, Dave.

Dave Vellante

>> Well, I know. And so-

>> I can't get CrowdStrike updated on my windows.

Dave Vellante

>> When we developed the power-law, obviously training was that the steep part of the power-law and the long tail was inference. And I'm glad you brought up agentic because I think a big part of that power-law, John, is going to be of reformation of the application stack, a new sets of intelligent data apps that emerge. So I wanted to come back to your Garman post and ask you if in fact this sort of came up, and specifically in order to enable that multiple agents to act in concert, you've got to have connections to data, particularly back end, even legacy infrastructure or legacy apps. So a key part of that is going to be Amazon or any company's ability to harmonize that data layer. So in other words, today you've got data in a database or a data lake. Amazon, as we all know, has multiple data stores. They've got technical metadata and operational metadata and business metadata all sort of distributed into the individual services. So it's not unified. They've got the ecosystem bringing in governance, and so you've got this mishmash of data pipelines which are very brittle and very hard to maintain. You're calling microservices. That has to change. So Amazon, I predict that Amazon at re:Invent is going to begin to unveil a unified data strategy that is going to enable those new applications, those intelligent data apps to be built. Did you get any indication from Matt Garman that they're moving in that direction?

>> Well, I tried to poke him on re:Invent. Of course he was tight-lipped, he knew what I was trying to do. But I clearly see that there's two things they're going to focus in on at re:Invent. I think they're going to go back, stay with their roots of working backwards from the customer, kind of that narrative of onto Cincinnati. When people ask about Gen AI, all the press, everyone's like, "Show me progress." They're beavering away. They're pedaling as fast as they can. It's clear there. But I think they're going to move the needle on two areas. Amazon and the cloud, and Amazon specifically, will move the needle in two areas. One, staying true to their bread and butter. Doing the muck work, the undifferentiated heavy lifting. So if you look at Andy Jassy's Q post around coding, it's about code refreshes. It wasn't saying, "Hey, we're going to replace humans." He basically says, "We just basically refreshed our platform on Java with no humans." What he's basically saying is all that shit work was done by agents. So you'll see that classic, mundane, undifferentiated heavy lifting. So I think that's phase one. Agents will do that. And then two, I believe this Tom Brady moment is Gen AI and I think they're going to look at developers. And I think Garman, what came across with Garman was is that he said, "Developers are the lifeblood of our business." He didn't say enterprise, he said developers. Now he didn't block the enterprise, but he was using developers as the high order bit in his mind of the value. And that applies to enterprises and startups. So I think Garman was clear that they're going to stay with their roots. He was very Amazonian in his cult like view, almost like Jassy. Again, Belichick, Parcells, he's going to be good. I think he's going to be solid, because look it, he knows the bread and butter. So I think on re:Invent you're going to see customer examples, their normal stuff, but they got to move the needle on the applications. They got to enable developers to build scalable apps and take advantage of the horsepower that they got with the infrastructure. If they don't do that, then it's just going to be a dud.

Dave Vellante

>> I mean working backwards from the customers, in my mind, had always been very focused on customer pain. What problems can we solve like today, and what services can we introduce to solve those problems? So very tactical. I do think some zoom out strategic thinking with customers is necessary here where customers are saying, "Look, we're doing all this Gen AI stuff, we're really not getting a big return. A big part of that is our data estate is all over the place. You've got to help us bring together our data." And I think that's what they did with, again, it was sort of a tactical hack with Zero-ETL when they announced that a couple of years ago. They did that when they separated compute from storage, it was really kind of a tiering thing, but they addressed those problems and they come up with solutions that work. And so I'm really interested to see what they do at re:Invent in this regard.

>> To use the Tom Brady example, when Tom Brady started, he took over for Drew Bledsoe on that injury, famous moment, and Bledsoe never saw another snap. Brady won games. Amazon's got to start winning some games. They got to start putting some real stuff on the scoreboard, real movement on Q, on other things. And if you look at where they have leverage on their game planning is I would look at their ecosystem, it's been in their core strength, they could really drop a major dime into the ecosystem's hands right now in saying, "Look it, here's some profit monetization opportunities." That lever could move the needle. And Microsoft's vulnerable, especially with the CrowdStrike debacle, that they're an antiquated kind of product compared to Amazon, but they got a killer ecosystem. Microsoft's Go-To-Market is classic enterprise and that's always been an Achilles heel for AWS, was how do you crack Microsoft's salesmanship and the way they got the grasp on the customers? Microsoft has that grasp. I mean, shit, they roll out Teams. Now Amazon has Chime, they don't really sell that, but I mean Microsoft just dictates to customers what to do. So in the cloud game, that doesn't fly 'cause cloud breaks as we saw with CrowdStrike. So Amazon has the opportunity to use that leverage. So put some numbers on the board and then ultimately win the Super Bowl, which is actually continue to be number one-

Dave Vellante

>> Which by the way, it's not true that Bledsoe never saw another snap. Remember when Brady got hurt against Pittsburgh, Bledsoe came in and saved the day. My point is he was a bridge to the future. So the core Amazon IaaS services can be the bridge to that Gen AI future, because they can perform, they can continue. To your point, people are running their business on Amazon so they can continue to generate revenue from that, and that can be a Bledsoe-like bridge, to continue with the analogy, to the future.

>> Well, my point too also was to your thing about Q. If you look at Amazon, one of the things they got criticized in the later part of their growth years was launching and abandoning things. So one of the hallmarks of AWS, we've been to all the re:Invents except for the original one, it was small, is that they do a lot of announcements, it was their calling card. Look at where all those new announcements, innovation. One of the criticisms has been that they launch and then it gets out there and they don't nurture it. And so if you look at all the game-changing needle-moving products from AWS, it's been DynamoDB, it's the database, these key elements. DCT obviously was the original one, part of the key building blocks. Amazon's got to roll out winning product that's going to be as game-changing as DynamoDB, or something like of Q, they got to make it work. You got to have new building blocks for this Gen AI era. I think that is going to be critical.

Dave Vellante

>> So one of the things you told me, I'm going to bring it back to VMware Explore, one of the things you told me is that he brought up Nitro. Garman brought up Nitro, and we've written about this and researched it, how it's a secret weapon of Amazon. And many of you might remember Project Monterey. Project Monterey was initiated several years ago and was an ecosystem-initiated development by VMware to develop a silicon capability, and again, working with ecosystem partners, to essentially be their Nitro. Nitro was a sort of a virtualization and network interface capability that allows silicon diversity. It has certain security and confidential computing capabilities to allow you to fence off access to the center of the core of the virtualization engine. And it allows silicon diversity, so we can accept Arm, we can do Intel, we can do AMD, and ultimately in theory would set up custom silicon. I haven't heard a peep about that project. I think it's dead.

>> Nitro?

Dave Vellante

>> No. No, Project Monterey. Right? Okay. Broadcom probably said, "Nah."

>> We'll take that on the chip side.

Dave Vellante

>> Exactly, that's what I was saying. I'm glad you went there. 'Cause maybe what's going to happen is the companies that were going to benefit from that, Dell, HPE, to a certain extent IBM, Cisco and others, NetApp, Pure, potentially they'll start working, Broadcom will start working with the Dells of the world or the HPEs of the world to actually develop that capability. I mean if I were Dell, I would be seriously looking at that because I think that's a gap that they have to... Relative from a cost standpoint, relative to the cloud. Now the cloud is expensive for many workloads as we know, but from a cost standpoint that gives the cloud vendors an opportunity and an advantage. AWS has it, Google has it. Microsoft has some similar even though they're behind. And so that is a capability that this ecosystem needs.

>> Well, I'm doing a story on this. I'm glad you brought that up and it just ties to the Broadcom chip side. One of the things I asked Garman was, "AI's really hungry for horsepower." Michael Dell had that quote like, "AI storage eats up AI," whatever he said. But I wanted to bring it back to the supercomputing democratization, and this has come up with Charlie Kawwas at Broadcom as well as Jensen Huang at NVIDIA. The democratization of supercomputing is happening, Garman did agree with that, but I wanted to ask him if I'm developers, are there ecosystems developing on top of the hardware and are people writing to the kernel level? You know what he said? Absolutely. He validated my statement, our thesis on theCUBE, again, another CUBE first. You're seeing software developers going down to the hardware. So then I asked him about this trend. He goes, and he put up the Nitro control, he goes, "Remember," I'm going to quote him directly here, Dave, I actually have the quote. "Actually, if you remember, our first chip that we delivered was actually the Nitro controller and thinking about the whole Nitro set." So it wasn't even a viable customer chip, it was a chip for the system. And we quickly, long ago, more than a decade, realized we first had the Annapurna team come join us. We realized building custom silicon in order to build custom network and hypervisor virtualization layer was incredibly important, over a decade ago. And that's how we get line rate networking. It's how we get better performance and storage. We virtualize all those into our Nitro system. Much better performance, super interesting, and that's how it's become a security layer because they can do stuff at line rate. That was the key. It was just a small little chip for the system. And the Annapurna team, you said it, one of the best acquisitions ever in the industry.

Dave Vellante

>> A $350 million acquisition.

>> It was super.

Dave Vellante

>> It's up there with VMware and EMC.

>> It's up there as one of the most strategic, genius moves, but that actually generated that hypervisor virtualization layer integrating into the chip level. And then he went on and on. Then he actually, he basically trashed Infiniband.

Dave Vellante

>> What'd he say about Infiniband?

>> Some need the most bandwidth, lowest latency and those between nodes and things like Infiniband, which is quite popular. Some of the other clouds kind of glommed onto it using Infiniband. But the problems of Infiniband, it doesn't scale at the rate you need. "Infiniband doesn't scale at the rate you need." Matt Garman, CEO of Amazon Web Services, said that.

Dave Vellante

>> And that's why-

>> It's very fragile. It's expensive. It's not like so long ago we've invested into networking that was ethernet.

Dave Vellante

>> And that's why-

>> There it is.

Dave Vellante

>> And that's why, by the way, Broadcom's all over ethernet with the Ultra Ethernet Consortium. And by the way, NVIDIA is also doing ethernet. So NVIDIA's bringing the CUDA stack and its software expertise to ethernet, so they understand that problem. You remember at GTC this year, Jensen basically described what they had to do, these unnatural acts to get NVLink working with Blackwell and these clusters. And so it's something that as an industry trend, I mean ethernet is unstoppable.

>> Here's the quote on this. Here's the quote on ethernet. "They're going to need reliability and be operable, and Infiniband just doesn't have those characteristics. Ethernet, we saw there was work to do. We made huge investment. We invested over the last decade. We get better performance out of our ethernet networks than you can on large Infiniband network. And those investments have things like EFA and our own protocols to help us get low latency, high throughput performance that customers need inside the cloud world." This is huge endorsement.

Dave Vellante

>> I think that's probably true at distance. I'm not so sure it's true at very short distances.

>> Yeah, I mean I think Infiniband, as I always said, you're close, but it's not your silver bullet. Ethernet is open, it's standard. You know how I feel about that. But I just thought that was interesting that Garman went in the weeds. I mean, I like this guy. You know why I like Matt Garman, he's a helicopter. He can go high level and go low level. He can sit with product guys and talk about engineering and then sit in a boardroom and talk about digital transformation. So I'm long on Amazon as I am with Microsoft, but I think Microsoft's got some challenges. All good, Dave, all good.

Dave Vellante

>> So the big thing everybody's waiting for this week, John, is NVIDIA earnings. NVIDIA And CrowdStrike both announce on Wednesday. NVIDIA, NVIDIA's up. Remember it crashed a month ago. It's up like 30% in the last three weeks, right? I mean people are piling back in. I mean, it's kind of gone sideways for a while, maybe it's kind of recoiling for either to go up or down, but that's going to be the big news this week. Everybody's watching NVIDIA, they're watching the demand picture. Are they going to make statements around Blackwell availability? We debated this on theCUBE Pod a while ago, you were talking about the popping the bubble, if you will. I still think demand, because of the CapEx spend, still outstrips supply. There was an article in the Wall Street Journal today about a reach-around, that Chinese companies are doing a reach-around to get H100s. You mentioned that up front. I think, look, NVIDIA, we forecasted NVIDIA's going to be well over $150 billion revenue company by 2028. If you look at the forecast for NVIDIA, revenue is doubling in 2024, up 40% in 2025, going down to 15% in 2026. I think they're going to potentially beat those numbers. I think they could potentially, John, be a $200 billion company as we enter the end of this decade. And so I still like the NVIDIA thesis, the long-term thesis. I know everybody's coming after them. I know there's a lot of talk about inference, but to your point, training and inference go together. I think NVIDIA has a lead that is maybe not insurmountable, but I think it's sustainable for at least five, potentially 10 years.

>> When you see companies like this that you say, "Oh, that's too overpriced. It's way overvalued." Every single successful company, you go back 30 years, Microsoft in the '80s, "Oh, that's too high. It can't get any higher than that." Facebook, "Can never get any higher than that." Google, all the top high-flying companies. The only thing that could hurt NVIDIA is themselves. And I think right now I believe their lead is huge. I think the mode is software. They're still a chip company, they're still priced by the chip, but their value's software. So as long as there's no major "world event," force majeure, something bad happens to NVIDIA, things blow up, allocation happens on the hardware side, they got this. Now the question on competition's going to be can they ramp up? So that assumes that NVIDIA's staying still. And this is what I tell everybody. Yes, NVIDIA's not standing still, they're investing. And I think that Jensen's so smart, he gamed this. He only went public knowing he has stuff in his back pocket coming. So if Blackwell drags on and CUDA doesn't really deliver or the customers aren't ready, that's the only thing that could hurt them. And I think, I am not a stock market buyer so I always cringe, "Oh, it can't go higher." It just seems like, oh my gosh, they're going to do a stock split, it's going to be back up to a thousand again. Oh, Dave, how many trillion dollar market cap is that at that point?

Dave Vellante

>> They're up over 3 trillion. They're banging the door at 3.2 trillion now. So what's next? Four? And five? I mean, I-

>> I know. Silicon Valley's going to be littered with NVIDIA millionaires now. And that's going to replace the Facebook millionaires, who replaced the Google millionaires. So another Silicon Valley trillion dollar baby, Dave. It happens. So it's like don't even think about buying a house in Palo. Alto.

Dave Vellante

>> You see the article? There was an article in the Wall Street Journal, I think it was this weekend or last week, about all these investors, these Main Street investors who put their money into NVIDIA. And there was an orthodontist who was obviously pretty wealthy, does well. He bought the stock, he put in like 60 grand, 50 or 60 grand in 1999, and he now owns $15 million worth of NVIDIA stock he has never sold.

>> NVIDIA since 2019, I should say, since 2019, jumped 3,776%. Okay? This is a headline on Bloomberg. Current and former NVIDIA staff say that despite the 3,776% stock jumps, minting many millionaires, the hours are just as grueling and high stress. Now, I remember Jensen said in that Stanford talk, basically what he said, you got to have a face point, you got to fail. He's like, look it, I mean Jensen, he's a workhorse. He's like a tech guy. He's like the Wayne Gretzky of tech. He's just a legend. So they're not resting on their laurels, to my point. I think the catch-up of the competition is just as important as watching NVIDIA's ability to execute continually. So it's like you're a horse racer. Remember Secretariat?

Dave Vellante

>> Of course.

>> Once it had the lead, no one ever caught it. So once it got from behind, like NVIDIA did, I think the competition question is how fast can they catch up to them and what do they gain? And how fast can NVIDIA not gain?

Dave Vellante

>> So just by comparison, so just doing some searching, and I'm not sure these numbers are exact, but according to Perplexity and Google, NVIDIA's forward P/E ratio is in the 70s today, and Cisco at the height of its P/E during the dotcom was 472 price to earnings ratio. So people often like to compare the dotcom to the current situation. I've written about this. There are many similarities, don't get me wrong, but that's not one of them. So one could argue, well, wow, maybe NVIDIA's not overvalued, maybe it's actually relatively cheap. If they're going to blow through these numbers, if they're going to be a $200 billion revenue company that's going to throw off... Maybe their gross margins will moderate a little bit, but they're still, if they're getting 60, 65, 70% gross margin, that might be an undervalued stock, John. Right?

>> Again, I've been digging into this 'cause I want to know. I've been trying to crack this puzzle, is NVIDIA overvalued or not? And I'm like the two-armed lawyer, on one hand and then on the other hand. So there's so much going on around it, it's just incredible. So I mean, first of all, I love the market. Now where I think NVIDIA could shoot themselves in the foot, if you had to pick one area, is overextending where they don't have leverage. That's why I like the HPE deal. But the leverage there is that if the HPE sales force can't handle NVIDIA sales, are they trained? Is the HPE channel and sales force trained up? Again, it's like playing Major League Baseball. Can they handle it, Dave? So I think they could, but NVIDIA's banking on that. And I love NVLink for NVIDIA. I'm not sold on the NIMS, 'cause I think RAG is popular now. But the agent post that you wrote, agentic post points out that there's life behind RAG. And again, two killer apps in my mind right now for AI is retrieval augmentation generation, which is vector embed neural network format, and alternate agentic services.

Dave Vellante

>> Let me ask you a question. So the forward P/E ratio on NVIDIA is now about low 40s, and the average P/E ratio for companies in the NASDAQ is about 25. So in your mind, given the TAM, given its market leadership, given its technology investments, given its leadership, is it justifiable that NVIDIA's forward P/E ratio should be under 2X the average of NASDAQ? I would say absolutely, 'cause if you think about Intel and Microsoft during its monopoly days, and I think NVIDIA has a form of monopoly, they were able to command a much more significant premium. I don't know exactly what it was, but I bet you it was like 4 or 5X over the average NASDAQ stock because of their market position, because they're essentially they're monopoly power. And I think it's hard to argue that NVIDIA doesn't have a monopoly right now. If you want the chips, you want the latest and greatest, you got to go to NVIDIA.

>> NVIDIA owns the market right now. They have a default monopoly by having the better product. And we are in a product-led and growth market right now. And if you look at all the companies that we cover, whether it's IBM, HPE, Dell, Oracle, all the way across the gamut of cloud, up and down the power-law, whatever you want to do, if you do not have a product-led strategy in this market's transition, in my opinion, you're toast. Okay, I'll repeat that again. If you're in this market and you don't have a product-led winner, that's transformative, you're toast. You're a me too player. Again, there's two sides of history here. Which side of the street will you be on? And ultimately this is where it's going on. And a lot of everyone's focused on this and it's trickling down into the trenches, Dave. The platform engineers, the database guys. If you look at the entire stack, hardware, middleware, apps, just to oversimplify it, every single part of the stack has been going through change. Every theater, business tech is under siege with change, and so it's opportunity.

Dave Vellante

>> The other killer business model out there that we haven't talked about in the context of silicon and hardware is Broadcom. They're not competing head on with NVIDIA and GPUs, but they are competing in terms of being the interconnectivity layer. So there's only a couple of companies in the world who can both do networking on the chip and networking at distance. And that's NVIDIA and that's Broadcom. And so Broadcom, because of its strategy from a silicon standpoint of betting on connect centricity, we heard this from Charlie Kawwas in our interviews with him, that is a form of monopoly in that they've locked up the market for the best silicon that connects all this stuff. And they've got Google, they've got Meta and they've got ByteDance as long-term customers. Floyer and I, we've got Broadcom blowing through 50 billion this decade, being a 55 to $60 billion revenue company. That's just in silicon. That doesn't include the VMware piece, or the software piece.

>> I think VMware is a cyclical thing that they can hedge. They got to get their act together, fix this quickly, get this revenue back up and durable and stable. But I will tell you that the power-law, again, we'll go back to the power-law, you're going to have GPU and connectivity challenges around the new custom silicon, the kind of horsepower that powers supercomputing. But here's the thing, the Dells of the world who have the AI factory, Michael Dell's vision is beautiful because what's happening is you got to have the power-law on size of device. And when you look at Matt Garman's, my interview with him, he says some things about inference. You have to have devices that have hardware that's performant based on the form factor and power it draws. So you have the big data centers, you're going to have mega power, and then you're going to see that get spread out distributively across that power-law where you're going to have the size of the giant footprints. So if you've got a little camera, you're doing inference, you're not going to do a lot of training there, you're going to send that back for training or maybe co-located somehow. So the data strategy's got to be different. And so you see a completely different architecture for that. Now, I'm a big believer in the AI PC is the right direction because all hardware is going to be refactored to be compute. So I think compute's going to be massively big opportunity or XPU or a combination of PUs like GPU, CPU-

Dave Vellante

>> NPU....

>> FPG&A. I think you're going to see-

Dave Vellante

>> Neural processing unit....

>> you're going to see a whole systems architecture that was once a motherboard exercise, happened into the computing realm. And again, when I talk about inference, this happened with Garman, he brought it up. He says to me, "Why would you want to have all that training done and not have inference?" Right? So he says inference is going to be huge.

Dave Vellante

>> Well, I mean that's the point of the training is so you can do inference. By the way, Dell also reports this week on Thursday. Some of the things you should look for at Dell, you mentioned AI PCs. People want to see evidence of AI PC, at least getting some traction or it starts to hit the monetization, become a meaningful portion of Dell's forward guidance. We'll see. But the big thing that people are looking for from Dell is in AI servers. Remember last quarter, their server business grew, and I might be off here, it might be AI servers or servers, but the growth was like 42% year-on-year. But the problem was that profits were flat. So imagine that John, you grow revenue 40 plus percent and your operating profit is flat. So the stock tanked. When people saw that, they said, "Oh my god, Dell's not going to make money in AI servers." And Dell sort of messaged, "Look, right now we're winning market share. We're selling at good prices to seed the base with our partners and MSPs in particular. But going forward, when we start selling to enterprises, the margin's going to improve because they're going to be more onesie-twosie." We'll see. The second thing I'm looking for there is what's happening in storage. Dell has been a shared donor in storage, and that is against their business model of gaining share on a consistent basis. So they've been donating share, and so we'll be looking for those two things. The other companies I'd love to talk about at some point is CrowdStrike announcing on Wednesday.

>> Yeah. I mean, I think we kind of run out of time here, but I want to quickly say that I'm not sure HP has the PC AI. I haven't heard anything from them on the AI PC.

Dave Vellante

>> Well, they've announced something there, right? So...

>> I mean, it's like they're hiding the ball. I think Dell is much more focused, even with the layoffs, I think they're going to go all in on AI. I think Dell's one of those companies that transform. You brought this up on a cue and we had a big debate about this. Internet, they transition. Michael Dell has a good eye for market transitions, and we always say, you win on transitions. They effectively transit from mail order to the internet and web, web to cloud and now cloud to AI. And I think they will lay off their employees and do that and get it done, and they'll come back stronger.

Dave Vellante

>> Layoff is, again, we debated about this, but to me the layoff was a Jack Welch move.

>> It's a lot of people.

Dave Vellante

>> Let's just pay. We got too many people, we're going to be more efficient, we're going to reorganize. We're going to drive AI into our business and automation and we can cut 10% .

>> And make the transition.

Dave Vellante

>> Yeah, absolutely.

>> Versus go out of business. Slow death.

Dave Vellante

>> Yeah. Or hang on to folks that aren't top 50%. .

>> We have a couple minutes left. Give us your take on the CrowdStrike and then the M&A stuff.

Dave Vellante

>> So CrowdStrike, the action in CrowdStrike stock has been really interesting. I mean, there's some data here. They peaked in July at around, the stock price, at 398. It dropped down on August 2nd to 217, so that's a 45% drop roughly. Now it's a 268, so it's up about 23% since August 2nd. So this month it's done very well. I love CrowdStrike. I mean, I have great respect for the company. They got a great product. They're CEO, George Kurtz, is a baller. Having said all that, the problem that they had occurred in July, John, and the biggest selling month of your quarter is the last quarter of the month. So obviously that hurt them. We heard Nikesh Arora in their quarter said, "We had to scramble to keep our quarter intact because of all the CrowdStrike distraction. Many of our customers were dealing with the blue screen of death. So we had to work hard to get their attention and close business in our quarter." I know Palo Alto had a great quarter, so they're clearly back on track. What I think is going to happen, I think George Kurtz is going to reset the quarter. I think the last quarter, everybody knows it's not going to be a good quarter, so that's why the stock dropped. The big question is what's the guidance going to look like? Of course, that sounds like bromide, but I think George Kurtz is going to use this opportunity to say, "Look, we're still working through this. We are going to reset our ARR targets way down." He's going to dial them down. I think people are going to freak out and say, "Okay, that's going to drive earnings down," and they're going to reset. So I think that's going to happen. I think they've still got this overhang of people are going to sue them. The ETR data that just came out, very clearly, customers are still in negotiation. They're giving discounts. They're trying to get renewals at a better price, and they're still negotiating on is there any compensation involved? We know Delta Air Lines is going to try to get a pound of flesh. So all that is going to cause, I think, a CrowdStrike reset. So I think you got to wait for that and maybe look for opportunities to jump in after that. I think it's a very dangerous game to be buying ahead of earnings generally, but especially here, it's very tough to predict. I don't like to do that. I know some traders that like to buy ahead of earnings, like I'm going to try to get in now. I think it's really hard to predict right now in this market transition.

>> I mean, it's like as they say, it's trying to catch a falling knife, you don't know where it's going to be falling or whatever. I think they're going to reset. I mean, they're going to have to take their medicine. And I think their product leadership is just too good and they don't go away. And same with Microsoft, even though the blue screen of death, which was... That's from the '90s. 30 years ago was the launch of Windows 95. There are still Windows 95 machines powering some manufacturing lines. Why? 'Cause the app works on it. I shouldn't say that, Windows workers maybe, but that's still a few years later. My point is, if you're running your business on Windows, you risk the Delta Air Lines situation. So I think people looking at their systems saying, if Microsoft can't be a core system where Linux and other more hardened open standards work, I'm not saying Microsoft's going to go away, I'm just saying Windows, it's not designed to be running an airline.

Dave Vellante

>> A lot of people are running their business on Windows though, right?

>> I'm just saying, Windows is not built to run an airline, as you can see from CrowdStrike. I mean, CrowdStrike fucked up letting it go. They own that. But they also point out the fact that on the Microsoft side it was a house of cards. So that means you can't run your business on Microsoft. That's the takeaway.

Dave Vellante

>> Well, it's a combination of Microsoft and CrowdStrike. Nikesh Arora, again, dealt with this. He said, "Look, first of all," he was very respectful, he said, "This is not a good thing for the industry, so we're not going to..." They're not going to ambulance chase, so I respect that. They compete with CrowdStrike, sometimes very vehemently. But basically he said, "Look, we do things differently. When we roll stuff out, we roll it out to a small 1% and then we go from there."

>> I know, but I'm not-

Dave Vellante

>> But CrowdStrike has a different philosophy. CrowdStrike's like, we got to go fast, we got to-

>> Yeah. Update faster. They're doing their job.

Dave Vellante

>> Exactly, because the adversary is fast. So if we roll out 1%, the adversary sees where the holes are, they're going to attack. So CrowdStrike's philosophy is speed, speed, speed. And look, this time it bit them in the ass, but it's going to be interesting. I'm going to ask George Kurtz at CrowdStrike, are you going to change your rollout strategy? How are you going to change that? I'm also going to ask him if anybody got fired.

>> Yeah, but they're at risk of the-

Dave Vellante

>> And you're saying it's Microsoft has some of the blame here.

>> No, I think that your line of question should be great. The question I would say to him is that he's at the will of the market, and the market is what the existing current situation is. The current situation is people are running their business on Windows. Now people say to me, John, why are you so anti-Microsoft, I'm not anti, I love Microsoft. I grew up on the Microsoft generation. I just think that... "Oh, you're pro Amazon." Yeah, they're like 200 times better than Microsoft at what they do.

Dave Vellante

>> At what?

>> Cloud. Scale. And so Linux is 10 times better than Windows and running shit. And what I am pro on is when you're re-architecting your enterprise, Microsoft was serving IT from the old generation. So you have this sprawl of all this IT with Windows and it's just not built for that generation, this new generation. Some of it's not multi-threading, it's not fully architected. So we are moving into a world where, as Garman points out, these things are integrated down to the high line speed optimization now at full scale. That's not Windows. So I think CrowdStrike was doing their job and they realized that Delta grew to a Windows shop. And guess what? When shit happens, you're out of business.

Dave Vellante

>> Well, they were doing their job, but-

>> So nobody wants that, by the way. So that's my point. If I'm in an enterprise and I am not asking the question to my team, what does our platform look like for the next 10 to 20 years? If the answer is Microsoft, I'd probably like, I want to look under the hood.

Dave Vellante

>> Well-

>> Like really?

Dave Vellante

>> You know what was fascinating to me, John?

>> Like show me the architecture that pens that together. If someone could show me how Windows is going to power a business, a multinational global company at scale versus say, Linux, fully managed, distributed, computing hard, big iron like stuff, that's why the mainframe argument's a good one coming back. So I'm just skeptical that that could happen. So that's why I like Amazon. I think Amazon, for all their good and bad, are high-scale supercomputing capability.

Dave Vellante

>> So ETR did a survey in July. They did it the day the problem occurred, the incident occurred. And of course there was a big emotional backlash. They just did another one that just closed just late last week. And look, the reality is it's not easy to move off of CrowdStrike. People, they have this emotional response, but people are still considering... There's no question they're changing their sort of thinking around this. But here's what's interesting. Which of the following vendors have you evaluated as a possible alternative to CrowdStrike? Number one was Microsoft. And I'm like, whoa, wait a minute.

>> Yeah, because it's hard to get off Microsoft too.

Dave Vellante

>> Okay, but you're going to move from CrowdStrike to Microsoft, where Microsoft was part of the problem and you're going to migrate to Microsoft?

>> Microsoft's argument is that we'll have better, tighter integration. So I think Microsoft will win on this. Again, it's just as hard to move off Microsoft.

Dave Vellante

>> Security's not as good. It's a good enough approach.

>> Well, I would argue that.

Dave Vellante

>> But behind that was SentinelOne, then Fortinet, then Palo Alto, all about the same, and then so for Cisco Tanium. But the fact that Microsoft was number one, I mean to me that was a major call-out. Now this is, again, a small survey. It's only an N of a hundred, but still, I mean that would not be my first choice, John.

>> I mean, look it, the survey says it's hard to move off CrowdStrike. This is the problem you have when companies rely on the vendors. When you have the switching costs that have dramatic and grave consequences of moving disruption, it's hard to switch. So CrowdStrike alone is a good product. The switching costs are so high, Microsoft's is even worse. They're so nested into these enterprises, you can't just rip it out. You can't just turn a switch and replace it with Linux.

Dave Vellante

>> So 40% of the respondents said they've evaluated alternatives in the last 30 days, but only 17% said, yeah, but we were doing that anyway. 23% said, we've evaluated alternatives in response to the outages. So 60% said, no, we haven't. Now maybe they've been too busy. So I feel like CrowdStrike's pretty sticky. I think it's a leading product, and they messed up. And so I'm really curious as to... I mean, right now they're being very transparent about what happened and they're negotiating with customers about what they're going to do about it. What I'm trying to understand is, okay, going forward, how is this never going to happen again?

>> Yeah. And I think your point about the resetting earnings, this is a really good interesting point you brought up, because I think that's smart. We had what happened to CrowdStrike, but I mean I see massive growth coming with Gen AI, but what I don't know is what the consequence will be for the folks who aren't on the growth side. Like I said, look at some of the companies going out of business. There's already kind of the talk of AI bubble bursting. People can't meet those revenue numbers. And so in every bubble, the question is does it pop or the bubble get air out of it slowly? And I think, to me, definitely there's huge growth, but will there be consolidation before this massive inflection kick up to the growth side? So constantly evaluating the market right now, and I think the M&A market will tell us more, and that's why I'm so anti Lina Khan, because there has to be robust consolidation for exit and liquidity to feed the system. And on with that just get through the bubble and get to reality, because in every bubble, whether it's a dotcom, web 2.0, everything that was in the bubble actually happens. So it's just, don't get too excited.

Dave Vellante

>> A couple things there. One is we heard Liz Warren say, "Thank God for Lina Khan, we need to give her more power." I don't agree with that. I think three letter agencies have plenty of power, in my view. In terms of the bubble bursting, the question is, okay, will it be a buying opportunity for people? I think yes, it will be, because you have extremely profitable companies, cloud companies and companies like NVIDIA and Broadcom. If and when the bubble bursts, and it will, to me, those are outstanding buying opportunities, probably better than was in the dotcom, which was very unclear who was going to make it through.

>> Well, it was when they weren't making any money.

Dave Vellante

>> Right, but none of them were making any money at the time, whereas today they are. So maybe the bubble will be a soft burst, a soft landing for the bubble. But I mean to me it's not like a Cloudera situation where Cloudera, awesome company, they started the big data, but their business model was never like a monopoly-like business model, like an NVIDIA or a Broadcom is. And so even though they got it all started and they were the source of innovation for today's big data era, they weren't able to get through the knothole. And now they're a good company, but not the great company that they potentially could be. Well, Snowflake, you look at Snowflake, right? That was the hottest company on the planet. And now the LinkedIn posts that we're on this weekend where one of the VCs, one of your contacts put out, "Hey, they messed up." They missed buying DBT or Fivetran. They missed getting through that knothole and now they've got some real challenges. That was an epic LinkedIn thread that you shared.

>> I mean, I think there's a whole world. Just as we end this special CUBE Pod here at the VMworld, I just want to let you know I got a comment on my LinkedIn. So following this analogy, does that make Adam Selipsky, Pete Carroll of AWS? If so, should we watch where he turns up next? Could win the next Super Bowl. He got it right away. Oh, Garman even commented on there. All right, it looks like the football analogy resonated with the sports fans. Obviously I couldn't resist the NFL example because of Thursday Night Football. Dave, always great analysis. Again, this is a CUBE Pod special analysis deep dive in the industry right now. A lot going on around Broadcom, VMware, a lot going on in the cloud. Again, the relationship between the cloud and on-premise and edge is driving a lot of innovation and a lot of kind of calibration of all the existing stuff out there to prepare for this big wave. A lot of stuff going on at the hardware level, at the data level, and VMware is right in the middle. And so are we. Here in the CUBE, getting all the data, share that with you, putting it into our neural network. We'll be back with more live coverage after this short break.