Cloud Wars, Company Wars, and Innovating Through Change

How do you transform and innovate as your company grows and matures, even as the underlying technology platforms shift? Michael Dell of Dell Technologies joins a16z's Martin Casado, Marc Andreessen, and Sonal Choksi to discuss.

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Speaker 00

  How do you transform and innovate as your company grows and matures, even as the underlying technology platform shifts beneath you? And what does this look like when you're a private company versus a public one? In this episode from October 2021, Michael Dell, founder and CEO of Dell Technologies, and one of the longest standing founder CEOs in tech, joins A16Z general partner Martine Casado, A16Z co-founder Marc Andreessen, and host Sonal Choksi on the occasion of Michael's book, Play Nice But Win: A CEO's Journey From Founder to Leader. They debate the impacts of the cloud wars, how innovation happens when a company is private and when it's public, something Michael knows well, having taken Dell public to private to back to public again,

• 0:43Speaker 00

  whether you can actually play nice to win as a leader, and more.

• 0:47Sonal?

  Hi, everyone. Welcome to the A6 and Z podcast. I'm Sonal and today we have yet another of our special book launch episodes on the upcoming new book titled Play Nice to Win: A CEO's Journey From Founder to Leader. The book is by Michael Dell, founder and CEO of one of the world's largest tech companies, Dell Technologies. And while his book interweaves the journey of starting, transforming, and keeping one's company, we focus primarily on the how to keep and how to transform in this episode. We also discuss the challenges of being public while trying to innovate beyond the behind the scenes stories with activist investors. And we debate the underlying tech platform shifts from the death of the PC to the current cloud wars.

• 1:29Sonal?

  So also joining this episode are A6 and Z general partner Martine Casado, who is co-founder of Nicera, which was acquired by VMware. Martine's also written a ton about the trillion dollar paradox of cloud, which you can find at a6nz.com/costofcloud. And we have A6 and Z co-founder Mark Andreessen, who joins as well and plays devil's advocate throughout. We begin with the cloud wars, then the company wars, and finally end on leadership. For context, Dell's company famously went public, then private, and then public again. And we also reference the write-up that we did here years ago called Making Sense of Dell Plus EMC Plus VMware in 2015. There were many analyses of that deal at the time. Through the tech lens, and there were many analyses of the complicated financing of the transaction,

• 2:14Sonal?

  but we were one of the very few who dug into and analyzed boat threads together to reveal a far more interesting story. You can find that at a6nz.com/delldeal. As a reminder, given all the historical events and companies mentioned, that none of the following should be taken as legal investment, business, or tax advice. For more important details, please see a6nz.com/disclosures.

• 2:37Martin Casado?

  So, Michael, when I was reading your book, by the way, which is fantastic, I just loved it.

• 2:42Speaker 03

  Thank you.

• 2:43Martin Casado?

  You're going through your psychology, particularly your belief that PCs still had a lot of legs, and then everyone was wrong. And I was thinking, this is the most extreme form of the founders seeing truth when everybody else doesn't. Because normally, the founders believe something the world doesn't, but there isn't a lot of data, and there isn't a lot of experts because it's so new. But in this case, this is incredibly mature. There's a lot of data and everybody knew, and still you hold this contrarian view. And so I was just wondering, I feel in infrastructure, we're having a similar dilemma now, where there is this consensus view that all of infrastructure is going to be consolidated in an oligopoly of free companies, and somehow all the PCs are going to disappear in the cloud, etc.

• 3:28Martin Casado?

  And so do you feel this is a similar situation, different situation? Would love your views on that.

• 3:34Speaker 03

  So there are lots of different models that are emerging, but the idea that everything is going to one public cloud, you really don't see a lot of customers doing that. I wake up every day and I look at our daily order report from the previous day, and it sort of tells me, no, that's not happening. But I can tell you from my conversations with customers, they've sort of figured out it's not the public cloud or the private cloud, it's both. And even beyond that, it's really the edge. And infrastructure is going to continue to be distributed. Yeah, they like the consumption models, they like the AI and ML services that the public clouds are building out,

• 4:19Speaker 03

  but they're not going to all be beholden to one cloud. More and more customers are thinking about what's the right place for any given workload. You've got data sovereignty issues, security issues, certainly cost issues. And look, when I look at what is about to happen with all this embedded intelligence and distributed computing being even more distributed and real time, I don't think all that's going to be concentrated. There are estimates out there that say there will be more apps and more data at the edge, and it's not going to all come back to the center of the universe.

• 4:58Sonal?

  Our partner, Peter Levine, actually wrote a piece a number of years ago, which was called "The End of Cloud", his point being that all of the computing is moving to the edges, as you know, Michael, and what that means architecturally because of the processing having to happen closer to those nodes at the edge versus like going back to some centralized thing. In some ways, the end of cloud was kind of a controversial statement, at least for those not in the know. And then more recently, this idea of cloud services versus on-prem and this whole debate that Martine and Sarah have been writing about. It was actually probably one of our most controversial posts in the last five years. And so I would love to hear you guys, you know, debate some of these technological shifts.

• 5:42Martin Casado?

  It seems like IT goes in these kind of like collapse, expanse, collapse, expanse phases. You have like mainframes, was kind of a collapse, and you have this kind of expanse in the PCs. Then you've had like a bit of a collapsing that happened into data centers and cloud. And now it's so clear we're in an expansion phase. Like there's a Cambrian explosion around data, faster connectivity, edge compute, edge devices. And every time you have one of these collapse expanses, the TAM grows by at least a factor of 10, only a factor of 100. So it seems to me like the TAM is increasing and a lot of this compute is exploding out of the clouds right now.

• 6:18Speaker 03

  Another way of thinking about this is, you know, the domain of technology used to be the IT department. And now it's the whole organization and the whole company. Basically, everything you want to do, from sales, marketing, product development, et cetera, anything in any organization, technology is the fulcrum. And that is driving a tremendous amount of demand and obviously opportunities for the industry. All of these things have to be built on a foundation of some kind of infrastructure. I mean, clouds are infrastructure and AI is eating software, software is eating hardware, but it's got to run on something, right?

• 6:58Martin Casado?

  I actually think it's instructive to go through the historical arguments quickly on this one because like the actual argument has shifted. So the first cloud wars, which I think most people remember, was literally like the cloud versus internal IT. And internal IT had old technology, you know, data centers were like really wiring closets and they were like museums of computer science past. And you didn't have automated processes, et cetera. And so traditionally, we're talking like when I was a programmer back in the early 2000s, there was a lot that you continued to reform that. And it turned out that having an alternative that somebody else managed helped. And so I would call that cloud wars 1.0. But in that time, infrastructure has evolved so much that that's actually not even really

• 7:40Martin Casado?

  a problem anymore, even if you're running your own IT. So that continues to play out, but something else has happened, which I think is even more significant, which is we've gone from basically every company is a software company, every company is a SaaS company. So instead of shipping software, now we run software in SaaS. And the economics of the cloud, if you're SaaS, are just really poorly understood. I mean, in the last six years, 126 companies IPO that were predominantly cloud native, 126. This is during a bull market where debt was almost free. And so we don't even know what it looks like when the market starts kind of seeing what their margins look like, right? And so you could argue that there's a massive shift, you know, to SaaS. I would call this cloud wars 2.0. And now when it's SaaS, infrastructure becomes cogs.

• 8:25Martin Casado?

  It's very, very important to have choice and flexibility, which is why Michael Dell and I have been so aligned on this for so long.

• 8:31Marc Andreessen?

  So I'll argue the other side.

• 8:38Marc Andreessen?

  You guys are agreeing with each other way too much. I'll concede upfront, like, there's a bunch that I would probably agree on, but not admit it. So, kind of argument in three parts. So, argument number one is you guys have been kind of infrastructure systems guys your whole lives. Like, I've been a software guy my whole life. And like, I took all the hardware design courses and like, I kind of, you know, use the toothpicks to prop my eyes open. I just couldn't wait to go back to the compiler. And so, the last thing in the world I want to do is deal with any aspect of hardware if I can possibly with it. Like, I just want to stay as far away from it if I can. I want to be able to type things, you know, in an editor. You know, I want to basically press deploy, but I just want it to work. And as software is the world, as software becomes more important and more kind of comprehensively applicable in kind of every sphere of human life,

• 9:19Marc Andreessen?

  and as more and more people become software developers, just the sheer amount of demand for people who just want to write code and press deploy and run, and certainly not ever deal with the concept of a server or a this or a that or a private this or anything other than just press and go. Like, that market is going to just like continue to grow explosively. So, let me pause there and see what you guys think.

• 9:38Speaker 03

  Yeah, no, I agree with that trend. And that's why we're seeing tremendous growth in basically creating this developer-friendly environment that allows customers to deploy easily. I actually think the arguments are somewhat compatible, given the trends in how infrastructure is now being delivered on a consumption basis.

• 10:01Marc Andreessen?

  And then, yeah, I guess the corollary would be you would view it as a challenge, for example, for your company is the environment that you're providing as a private cloud has to be as easy to develop on, has to be as easy to sort of click and go as a public cloud. Is that the way to think about it?

• 10:14Speaker 03

  Yeah, I think one of the other things that we're seeing is the super big organizations, I think this will flow down to large and medium-sized organizations over time, have basically said, "I'm going to put my data in a neutral colo that has incredible connectivity, and I'm going to access services from each of the major public clouds, because it's just an untenable situation to put everything in one of these." And I actually think that colo model is growing faster than the discrete public cloud model itself.

• 10:55Martin Casado?

  And just to follow on this, you think, Michael, in line with that, what happened in the pharma industry is it turned out Big Pharma would rather erode their own market share than drop their margins. And so the buyers propped up this company that was just made for generics called Dr. Ready's that was like actually built for like 15% margins, et cetera. Do you think that, you know, to your colo model, which by the way, I agree it's a massive, massive trend, do you think we're going to see the rise of almost like a generics type cloud that uses merchant hardware in a similar model to that? Or do you think it's going to be every enterprise does their own thing?

• 11:27Speaker 03

  I mean, I've seen a number of business plans like that. It feels to me like it's going to be highly distributed, particularly when you think about it globally, you know, there are data sovereignty issues that are real and significant. And with the rise of these enormous tech companies, you're seeing some nationalism pushback on the usage, even if they're making significant investments inside those countries. So it feels like it's going to be highly distributed and that's kind of what I see.

• 11:59Marc Andreessen?

  So second argument is sort of the consequences of the rise of AI and machine learning. So Michael, well, we'll tell you what kind of what we think, which is the nature of software development is about to change dramatically and it's going to change from basically people writing deterministic code, you know, sort of the traditional style of like first this, then that, and the other thing, kind of in a classic programming model to writing much more software that's basically built on ML models. So basically software systems where you dump in huge amounts of data and neural nets do their thing and they they come out with answers on the other side. But of course, we're already seeing the rise of that kind of development, you know, large parts now of the big internet companies like Google and Apple now run on these systems. As an example, like Google has reengineered its entire search engine to run in this model, which is a real harbinger of the future.

• 12:42Marc Andreessen?

  And we're seeing it across, you know, many areas of tech, including, you know, drug discovery and tons of others. And so a couple of things happened as that shift happened. One is just like the whole nature of development changes because a programmer, you're dealing with software at kind of a different level of abstraction. And in fact, a lot of programmers who are going to be writing against AI are not going to understand the details of how AI works. They're going to basically be writing on these high level APIs that access these very complex algorithms. Kind of under the hood. And then the related thing is the demand for processing power is just going to explode stratospherically from here to right. Nvidia, you know, is a major beneficiary of this trend today. It's become hard, at least for a lot of consumers, to buy GPUs, even for their PCs, because those are the same chips that you want to run these AI algorithms. And so you just have this sort of massive explosion of the number of cores and

• 13:26Marc Andreessen?

  the number of processors and GPUs that are going to be necessary to support this, your programming model. And so I think the argument would be that in this model, then you really want a cloud, right? Which is you really want a super high scale environment with a great deal of flexibility to be able to scale your processing needs. And with an increasing amount of specialized hardware, in the form of these GPUs and all of the custom silicon that's coming along to support AI. And you really want higher and higher levels of abstraction for the software developers, which again involves lifting them further and further out of the nuts and bolts of how the infrastructure runs. And I'd be curious what you think of this trend for the kind of model that you're talking about.

• 13:59Speaker 03

  I think there are going to be workloads where the scale of the public clouds is going to be super efficient and going to work well. But I don't think that all the organizations in the world are going to say, hey, we're creating all this valuable data, and we're just going to give it to you, and then we'll rent it back at whatever price you want to charge us in the future, because we don't have any other options. Look, these companies are creating a lot of powerful services that customers want to take advantage of. But a lot of those AI ML frameworks are not just owned by those three companies or exclusive

• 14:42Speaker 03

  to their public cloud infrastructure. It's a lot of open source activity there and companies, you know, helping companies figure out how to use all their data, which we're still at the very beginning of that as an industry.

• 14:56Marc Andreessen?

  Are you seeing AI demand or machine learning demand?

• 14:59Speaker 03

  Yeah, absolutely. It's one of the fastest growing workloads that we're seeing. You know, certainly Nvidia driven and there are many others that are working in that space too. But that's driving enormous demand for sure.

• 15:13Marc Andreessen?

  Yeah, and then I think the third argument would be basically given the nature of competition, which is it was not that long ago that it was at least taken as an article of faith, maybe incorrectly if the cloud businesses were not going to be profitable or can be only marginally profitable. And obviously you guys and Martine with your work have done a great job of illustrating how actually it turns out they're actually quite profitable. But you do have this dynamic, you know, Amazon, Google, Microsoft. These are three now, just like, you know, gigantic companies with what seems to be a very high level of determination to bash each other's brains out in this market. They're all so determined to win this war. Aren't they going to beat each other to death on pricing over the next five or 10 years? And won't that result in their margins actually coming down? And therefore maybe resolving some of the economic pressure, Martine, that you wrote about?

• 15:52Martin Casado?

  Well, I mean, I just don't think oligopolies work that way. If you look at the actual margin profile over the last 10 years, it's gone up from about 18% to 30%, right? So it's actually increased. So while price is dropping, that's through internal efficiencies, margins have been increasing. And so I think whether it's being checked by the public markets and/or the internal operations of the company is a very, very hard thing to do for them to erode margin. And again, we have examples where you can look at, I think Big Pharma is one of the best ones where like they would just rather cede market share to generics than drop their own margins. Listen, if oligopolies change the way that they work, then maybe we're going to see an erosion of margins. So prices will drop 100%, but I don't think margins will. And as workloads become more predictable, so let's say you're a big SaaS company, you're a deck of billion dollar SaaS company.

• 16:38Martin Casado?

  Your margins are going to be very important as well. Your workloads are going to be predictable. Your workloads are going to be at scale. You're going to have huge teams that know how to run platforms. You would be silly not to take some of those workloads and you collect the margin gain as opposed to Amazon. You'd just be silly not to. And I would say one more thing is very interesting, which is When the big three decided to start disrupting the server supply chain, they did it when servers were 50% of their COGS. That was the point in time, it was about 2007. And if you look right now at these big SaaS companies, cloud is 50% of their COGS. So something's gonna happen.

• 17:12Speaker 03

  Yeah, it's all owned by three companies. So I just don't really see that scenario happening. Here we are in 2021 and server demand is off the charts strong. And if I look at the ship to addresses for those servers, it's not all to three companies. It's global. It's all these new kinds of workloads. And again, I think with low latency networks coming into play, it's likely to be even more distributed in the future.

• 17:44Sonal?

  So on the cloud pricing wars discussion that Martine, you and Michael were having, it struck me that you guys were actually talking about more mature companies versus like early stage companies where part of Mark's question, at least the way I interpreted it, is that it's also about basically how you get innovation, right? Because the whole point of a startup, an early stage startup, being able to go cloud is you don't have to worry about the plumbing and all the crap, like you can just focus on innovating. So can you guys talk a little bit about the tension between the early stage model and then the transition to mature here? And by the way, Michael, I think this is really relevant to your book because it takes us to the narrative where you're really talking about what it means to go from an early stage to a mature company and then how you continue to get innovation. So let's talk about this whole debate about margins

• 18:28Sonal?

  and what happens and the tension between innovating.

• 18:31Martin Casado?

  Yeah, for sure, sure. So, you know, the cloud has a ton of benefits, which is, you know, if you're starting a project, you can do it incrementally. You get the cost benefits of something that's, you know, already at scale. You get all of the latest kind of softwares and tools, et cetera. So if you're a new startup, it makes a ton of sense. And you're building a SaaS app or something like that. You're not for infrastructure. It makes a ton of sense to use the cloud. Or if you're, let's say, in a big company and doing a new project and the cloud already has some tooling that your internal IT doesn't, it totally makes sense to use the cloud. So there's a lot of value in the cloud, for sure. The issue is, and this is kind of where this argument is just a new argument, is we're no longer in the old enterprise world, right? It's not cloud wars 1.0, right? We're in this

• 19:12Martin Casado?

  world where most software is run as a service. And that's just a different world. It's a different world because you actually have people that are good at operations in your company. It's a different world because the infrastructure actually becomes part of your COGS. And so as you get larger, the implications become more significant. And so there's a bit of a tension or a paradox right now where early on, it totally makes sense to focus on using cloud workloads for some projects, but as you get larger, you're actually kind of reselling the cloud and you have to decide what to do about that. And so this is what the industry is grappling right now. And as I mentioned before, and I think this is something everybody should remember, 126 companies IPOed in the last

• 19:55Martin Casado?

  six years during a bull market with almost no debt. And so this is a very, very new phenomenon where you've got companies that have built entirely on the cloud that haven't had to worry about market pressures because they're growing a lot and they can basically raise money for free. As soon as that shifts, I think that all of them are going to have to really look at, what does it mean that 50% of their COGS is coming from one of three companies? It's a very big deal.

• 20:18Speaker 03

  Yeah, I agree. And the trend that we've seen is companies get to a certain size, they start to focus on this. What you also see is the startup ecosystem maturing where at some point in its journey sort of realizes, hey, wait a second, we're going to be like paying enormous rents to these public cloud guys and we're never going to be profitable given the cost that we're paying for infrastructure and for the egress fees and everything else. And so then they get into a more complex analysis of their workloads and they start picking off parts. They're like, wow, this is going to cost us way less than we're paying over here in the public cloud.

• 21:03Speaker 03

  For certain things. And look, I think the rise of Kubernetes is very much a function of this idea of people like clouds, they like cloud native software development, but they don't want to be locked into a particular infrastructure location. They start to architect to be independent of location. And we're seeing that. So it's real.

• 21:27Martin Casado?

  Yeah, the important thing is like, this isn't a categorical anything, right? It's much more about like building anything like you use whatever's best for purpose, right? And like, as part of this kind of constraint is going to be kind of costs and cogs. And so like, the more flexible the workloads are, and they're being built to be more flexible all the time, the more, to Mike's point, and I think he's right, they're going to be distributed. And that means that like, some of them are not going to land on the cloud.

• 21:51Sonal?

  I will leave the whole repatriation debate off the table. I'm just chuckling because I was remembering that phrase that someone shared, the rise of the repatriarchy or something ridiculous.

• 22:02Sonal?

  You know, along those lines, one of the things that really struck me, Michael, in your book, you talk about as a leader of a public company, how there's always this tension, and in your case, like between profit and growth and share. And do you think that actually those three would be so aligned? And that they should just all kind of go forward, lock step? Like, why doesn't profit lead to increased growth, to increased share? Can you talk a little bit about some of the tensions there?

• 22:27Speaker 03

  Sure. If you're going to invest heavily, it's going to drive your profits down. And so it's always this balance of the three. And you can drive share up pretty easily. That can lower profits in the near term. If you're going out and hiring a bunch of salespeople, investing in new accounts, building out new relationships. And so you, again, you have to find the right balance. Now, if you've got a great business model and fantastic margin structure, they can all align very nicely, but that's not the case in every scenario.

• 23:06Sonal?

  That was great. We're going to shift into talking about the Dell EMC. Story, you know, you went from taking a company public to private and then back to public again. Mark is the one who made me write like with the corporate development team, the Dell EMC primer we did, and I kind of hated him for it. I was like, what am I writing this weird finance thing? And what is the tracking stock? And I was so annoyed because I'd never written anything that was like corporate finance. I ended up coming out of it the other side, like in love with the whole topic. It was fascinating.

• 23:36Speaker 03

  You guys did a great job with that.

• 23:37Sonal?

  Thank you. Well, you guys gave us a lot of great material to work with. It was Mark's idea. He was the one who gave us the through line because I was like, Mark, what's the entire point of this? And it's all about how do you stay innovative in a market environment today? So tell us about that.

• 23:54Speaker 03

  Yeah, so it was quite a journey and you kind of have to go back to the mid 2000s when things really started to accelerate in terms of change in the industry. It became kind of clear to us that while we'd had a lot of success up to that point, we needed to do some different things. We need to invest in software and new capabilities. We were going to have to build and acquire a lot more IP. We started acquiring companies and the more we did that, the less the market really appreciated the change that was going on. That was kind of painful and made us sad, certainly because we were working really

• 24:37Speaker 03

  hard at it. But at some point, it became an enormous silver lining because it created this incredible opportunity to buy the whole company back from the shareholders and really accelerate the transformation with no quarterly shot clock. And so we did that. We just slammed on the accelerator and we just said, all right, there's no shot clock here. We're just going to hire tons of engineers and we're going to hire enormous numbers of salespeople and we're just going to go for it. And we did an enormous level of investment organically inside the business and it started to work.

• 25:20Speaker 03

  Pretty quickly. The go private was going really, really well. Cash flow was coming in. In fact, we were, within 18 months, our net debt was like zero, which was pretty amazing. So anyway, about a year after that, Egon Durbin from Silver Lake and I started talking about, what can we do next? We knew we needed to transform more. EMC had become the leader in data storage and they had this 81% interest in VMware. And we had this idea of, wow, if we could somehow combine Dell and EMC and VMware, you would create this just amazing company. The only problem is we needed $67 billion, right? And so we worked on that

• 26:05Speaker 03

  and through some very creative structuring, figured out a way to do it kind of with their money, basically, and contributing all of our own equity. And we thought it would work, but it actually turned out to be way better than we thought. So that's a quick snapshot of it.

• 26:23Sonal?

  Can you put a little bit more in context, like just a little bit more about the underlying tech shifts besides just the players and the financial part?

• 26:31Speaker 03

  Sure. I think the backdrop for a super long time has been this incredible growth in the amount of data. And that is only accelerating. But of course, you have to turn that data into something useful. And one of the things that was happening in infrastructure and clouds was this extrapolation of the offering up to a higher level where nobody really wanted to assemble all these components themselves. And so you had things like hyper-converged infrastructure, converged infrastructure, cloud. And if you were just a server company, or just a storage company, or just a networking company, you really would have a difficult time. The big thing there is customers were telling us

• 27:17Speaker 03

  They didn't want to be systems integrators. They didn't want to put 20 things together. And of course, all the other 20 things were not number one. We were number one in basically all of them, which made it super compelling and I wouldn't say easy, but it was very straightforward to go to these customers and bring them a much broader offering.

• 27:38Sonal?

  Great. Martine, we all heard from the market their reactions, but what was your reaction from the inside? I mean, you were at VMware at the time.

• 27:45Martin Casado?

  I was, yeah. So I was running the network and security business unit. I actually remember when this happened. And what was clear to everybody within VMware was that infrastructure was becoming a battle of titans. It was companies that owned the entire stack. It's like every single player was kind of wedded to something pretty serious. And so the only two really neutral players, it seemed, were Dell and VMware. So you have like the full kind of hardware side and the full software side. And so there's a tremendous amount of excitement because it wasn't super obvious how you could play a game this big without being part of a much, much larger offering. In order to stay super relevant, like something like this had to happen. But, but listen, it was also crazy scary because we didn't

• 28:27Martin Casado?

  even know what a tracking stock was. So there was kind of a lot of kind of education that happened along the way.

• 28:34Speaker 03

  You know, the other thing that was amazing was it was pretty rare that a number one business was ever for sale. And if it was for sale, was that at some ridiculous price, people were saying, oh, all infrastructure is going to go away because the cloud is here. And, you know, that turned out not to be exactly the case. It's multi-cloud. And I think the rise of the cloud also discounted VMware and EMC, just like earlier, people had said, you know, the PC was going to go away. There was this narrative that kind of enabled us to go private, which was At the time, sort of 2012, the smartphone was really on the ascendancy and the iPad was

• 29:17Speaker 03

  taking off and everybody was like, "Oh, you don't need PCs anymore because of these smartphones and tablets." And we always thought it was an "and" thing, not an "or" thing. So that demise of the PC narrative gave us this incredible opportunity to buy the company back. The GoPrivate started in 2012 and was completed in 2013. The announcement of the deal was in 2015. It closed in 2016. That was quite a not-hole to get through with Carl Icahn and other adventures along the way, which I describe in the book.

• 29:53Marc Andreessen?

  Yeah, so Michael, you and I have both had the rather rare sort of experience of tangling with Carl Icahn over the years. And I'm very used to arguing your side of things. Let me argue his side, try to steel man his side, and then see what your perspective is on this now that some time has passed. So his side of the argument, if I recall correctly, was basically, a sitting CEO of a public company buying his own company is the most obvious conflict of interest in the world because you have all of the inside information and then you have control of the process. And so you, in theory, as a CEO, can do all manner of things to manipulate how the company appears to be doing, what the price should be, what the start price is and so forth, and then the bidding process and the buy process itself. And so you're basically in a position to cut yourself an amazing deal. And I haven't heard him say this, but I would imagine, you know, he might

• 30:35Marc Andreessen?

  even say today, "Aha, I was right in retrospect. Like, look how well it went, close to buyout, and isn't that evidence that the original thesis was right?" So now that some time has passed, what is your perspective on that argument? And what's your perspective now on the role that activists play and your take on how that whole thing played out?

• 30:51Speaker 03

  Sure. So let's unpack that. First of all, I think if it works, it's easy to say in hindsight that the founder or the management team knew what they were doing and they got a great deal. But if it doesn't work, you wouldn't really say that, right? Of course, there were enormous protections that were put in place by the independent board and special committee that oversaw the process to ensure that the shareholders got the best deal possible. And in the end, Carl Icahn, for all of his bluster, and he went on TV and lied about all sorts of things, which I detail in the book, he never offered one penny more than Silverlake and I were willing to pay for the company.

• 31:37Speaker 03

  What I discovered in a dinner at his house over his wife's meatloaf was that he basically didn't know anything about the company, didn't care anything about the company, actually didn't even want to buy the company. He just wanted me to pay a little bit more so he could make a little bit of money and go on to his next adventure. For him, it was just a big poker game. I asked him what his plan was, and he sort of mumbled a bunch of things. So I sort of called his bluff and said, okay, well, you know, why don't you buy the company for a little bit more than I'm willing to pay, and you're going to totally screw it up, and I'll buy it from you for a lot less than you paid for it. You know, that'd be a great deal for me. And it was pretty clear he was scared of that scenario. And it was another

• 32:22Speaker 03

  interesting thing that maybe very few people were following this, but every single share that he bought up until that point was above the deal price. So if the deal had gone through, he would have lost money. Now, he later convinced Southeastern Asset Management, our largest outside shareholder, to sell him a bunch of their stake at a discounted price. How he actually did that, I don't know, but if you sleep with a tiger, sometimes you wake up and your arm is eaten off. Look, I don't have a problem with activist investors. I have a problem with green mailing and basically misleading investors, which is what he did.

• 33:05Speaker 03

  Specific example, he sort of claimed that you could claim appraisal rights, which is a very arcane aspect of go privates, and there was absolutely no risk. And he was saying he was going to do it, which he ended up not doing, and it was not no risk, and he kind of knew all those things, but that didn't stop him from going, you know, on national TV and line. And so that's Carl Icahn for you.

• 33:30Marc Andreessen?

  Right. I steel man his point of view, but like I said, I'm already starting on your side. So if I can steel man your point of view and tell me if you think this is right, I think it would be the ultimate expression of faith by a CEO in the business is wanting to buy the entire business. And so when the CEO stands up and says, I want to buy this thing and for this price, it should be a massive best signal to every other potential buyer in the world saying like, this thing is like really valuable. This thing really matters. The CEO really believes. And so therefore, anybody with the money to do it ought to like really pay attention to it and come in and try to bid more. And the process in fact supports that.

• 34:01Speaker 03

  Yeah, exactly. That's what happens when you start a go private. Effectively, the company is put up for sale, right? And so any number of things can happen at that point. You can have other bidders come in, the price of stock could go up. There could be, you know, various forms of recapitalizing the company such that a go private is either no longer practical or necessary. So all of that could have happened at any point. Any number of buyers could have come in and said they wanted to pay more than Silver Lake and I were willing to pay, but nobody did.

• 34:38Marc Andreessen?

  Was there ever a moment when you thought somebody else would come in and do that? And end up buying the company?

• 34:43Speaker 03

  I mean, Blackstone was hanging around the hoop and talked about offering, I think it was 14 or $14.25 a share. They ultimately backed out. All the private equity firms took a pretty good look, but all except Silver Lake got pretty scared by what was going on in the PC market at the time. Windows 8 was sort of a failure and it just played into that narrative of the smartphone and the tablet, but with operating systems, you have effectively hardware and software working together to create systems. You've got cycles that effectively create a

• 35:31Speaker 03

  Refresh and a new source of demand as you get new functionality. That's true with processors, with operating systems like Windows and others, and they enable new capability. We knew that it was just a matter of time before there was a better version of Windows. You know, these machines were getting four, five, six, seven, eight years old, and there would be enough new technology that would motivate people to replace them. I think a lot of the private equity firms basically just looked at the business as cells on a spreadsheet and didn't really understand things like operating system cycles and some of the changes that were going on deeply enough to be able to underwrite the business.

• 36:16Sonal?

  Mark asked you about lessons learned in fighting an activist. One of the things that really struck me from your book is the narrative side of that lesson in that you're not just fighting with facts. In fact, you can't fight with facts because there are many cases where you had media outlets completely misrepresenting the numbers, there were analysts misrepresenting the numbers. At one point in your book, you actually do a calculation that Carl Icahn, for claiming all of his, "I'm trying to get more shareholder value," actually underperformed against the S and P 500. And so the facts do not match the perceptions that are out there. So I'd love to hear your thoughts and kind of narrative lessons for founders there.

• 36:55Speaker 03

  Yeah, look, Carl Icahn's record speaks for itself. All you have to do is look up the stock of Icahn Enterprises and compare it to the S and P 500 over any period you want, but that doesn't really matter. He's a magnet for media, and so he gets views, he's interesting, he's provocative, he says stuff that causes people to continue to watch, and that's what they care about. So I think you have to be a good person and be someone that people trust and want to partner with and rely on. And when you're going through something like that, you're going to tap into, hopefully, all of that equity that you've created over time with everyone.

• 37:35Speaker 03

  You've got to tap into the relationships that you've built to keep customers and team members and everybody focused.

• 37:44Sonal?

  If you started a company today, Michael, do you think you would take it public sooner than later, or would you wait?

• 37:50Speaker 03

  You know, we went public in 1988 because we really didn't have any choice. That was the only way we could get capital to grow. I think you only want to go public if there's an objective reason to do it. Companies that go public too early, I think, are in for a bit of a rude awakening in terms of what they're subjected to. And you really have to be ready to go public. So I wouldn't be in a big hurry to go public.

• 38:18Sonal?

  You know, obviously, one of the big things that comes up is like the number one reason to go public is to do mergers and acquisitions. And that was a big part of your strategy in terms of thinking about how to kind of create the company you want. What was the tension in your head between the strategy of M and A and what you were able to do inside? Like, how did you navigate? What do you build versus what do you buy?

• 38:39Speaker 03

  Well, ours was kind of a unique situation because we did this massive acquisition. It was highly transformative for the capabilities and the position of the company. And I think if we had been a public company, I don't think we would have been able to do that. You know, we would have had to have votes by both shareholders and public shareholders on the acquiring side. You know, if we were public at the time, probably would not have accepted that kind of risk that, to me, looked like a great opportunity. So you have to understand where your co-investors are, whether it's public shareholders or private co-owners,

• 39:22Speaker 03

  and make decisions appropriately.

• 39:25Marc Andreessen?

  What advice, if any, would you give to the next CEO of a public company that's kind of in the position that you were in to potentially buy your own company? Is there advice you'd give them based on having had that experience now?

• 39:36Speaker 03

  Well, you better get yourself a really good lawyer and follow the proper procedure super carefully and be ready for something that is way harder than you could ever imagine. The public market doesn't like it when you try to take away their products, you know, the public stocks. And founders are assumed to know more than the average person when it comes to buying back a company. Fortunately, this one worked out, so I guess it proves the case.

• 40:05Martin Casado?

  So what I would love to dig into, Michael, is markets transition all the time. Transformations happen periodically and they almost always kill companies. It's very rare that leadership that gets brought in later can navigate a transformation. They can sometimes, but you know, often they don't. And I would love your thoughts on why you think that is. And then I was also wondering, it's kind of the second part of the question, which is every once in a while you have a leader and it's normally a founder that navigates transformation, right? Reed Hastings famously with Netflix and you've done it now multiple times. And you know, I was thinking you might be the last actual like of like the great founders standing. Is that the case?

• 40:43Speaker 03

  You know, Larry Ellison's still going, depending on how you count, but yeah.

• 40:46Martin Casado?

  Well, so this, you know, Jensen and Larry Ellison, the only two others I can think of, but I don't think they've got this level of involvement because like, you know, I've seen so many come and go in the last 30 years and I think you're it.

• 40:58Sonal?

  I have a quote here. I actually have a quote here. You're one of few remaining founders who's still CEO 36 years later. Like, that's amazing.

• 41:08Speaker 03

  Thirty-seven, yeah.

• 41:09Sonal?

  Thirty-seven.

• 41:09Martin Casado?

  I'm just so interested, like, why founders pull this off more than other CEOs? And you can believe it is because founders are special and so they see truth. And then you could also just believe that they've got more credibility because they're founders, they've got more juice. Do you have an opinion or maybe it's both, like?

• 41:25Speaker 03

  I think founders get some extra permission to drive change that you might not have as a non-founder. If a company is sort of doing okay, but it needs to transform, you have to create a real sense of crisis inside the company and kind of balance scaring and freaking people out where if we don't do this with inspiring and motivating them about what the future can be. And that's sometimes a tricky balance, but to motivate change and transformation, you have to strike that balance. If you screw it up, of course, you can lose all that

• 42:07Speaker 03

  trust and credibility. But if you are thoughtful and are straightforward with people about what we're doing and why we're doing it, they'll run through walls with you.

• 42:19Sonal?

  Okay. So the title of your book is Play Nice But Win. And the subtitle is a CEO's Journey From Founder to Leader. And you have this line that you said that you were never looking for a fight, but relishing every brawl once you're in it, which I think is just fabulous. Do you think that people can really play nice to win? I mean, obviously, it's a premise of your book, but do you watch Ted Lasso by any chance?

• 42:43Speaker 03

  You know, I just started watching it because so many of my friends have told me, you know, it's hilarious. So yeah, I think I'm into episode four or something like that.

• 42:52Sonal?

  Okay. So you're still on season one?

• 42:54Speaker 03

  Yeah, still on season one.

• 42:55Sonal?

  Well, I'm only asking because there's this whole now argument that people are making that we're in the Ted Lasso age of management, that managing by nice is the way that people are going to get ahead. But the thing that I find so funny about it is that the show itself is actually more entertaining because of all the assholes that are in it. And the fact that you have this juxtaposition of like Ray Kent against Ted Lasso. And so I wonder if you have thoughts on how to deal with the assholes in the workplace and this mind of play to win?

• 43:25Speaker 03

  So I'll tell you a little story. Every so often, I will get a group of people that are not on our executive leadership team, you know, maybe 10 or 12 people that could be the future executive leadership team. And I'll spend three or four days with them, give them each pretty big assignments. They share their work with everybody else. And I kind of see how these people would work together with me as if they were the leadership team of the company. And every once in a while, you'll see somebody in there and say, okay, this person's not helping the other people sort of only out for themselves.

• 44:09Speaker 03

  And thank you very much. You'll be going to work somewhere else. And that's just how we do it. And by the way, this person had a really big job at our company before that. So I won't say who it is, obviously.

• 44:22Marc Andreessen?

  What's his name? Martin Casado?

• 44:25Speaker 03

  No, it wasn't. It wasn't Martin. Wasn't Martin. I can confirm that.

• 44:29Marc Andreessen?

  Yeah, clarify that upfront. So now your denial is suspicious.

• 44:35Sonal?

  Martin, what do you think about this whole thread?

• 44:38Martin Casado?

  You know, it actually reminded me of something, which is, um, you know, I've been in the industry for a while. And when it comes to reputation, Michael, I think you've managed probably to have like the cleanness as far as like being fair, being direct, being well respected. And you've done it for a very long time. And I said, actually, honestly, industry wide. And I'm just wondering, like, is this something that you've intentionally worked on? Or is it just kind of a natural byproduct of how you kind of navigate yourself? Because I do think that people's reputations really matter in business. I think business actually follows behavior. And I think you're one of the people that should model themselves after, so.

• 45:15Speaker 03

  Well, thank you for that. I can't claim that I actually did it intentionally. I just woke up every day and did what I thought was the right thing to do.

• 45:24Sonal?

  A quick question for Mark and Michael. Michael, do you have any regrets about leaving college, you know, to start your business? Because I think you said you did it before you were 20. And I wonder, because Mark is always the one who critiques this whole notion that people should leave college.

• 45:37Marc Andreessen?

  Would you recommend to people that they leave college to start a business?

• 45:40Speaker 03

  You know, it's not for everybody. I mean, you have to be kind of a deviant and mischievous and a rule breaker. And that's clearly not for everybody. And it doesn't always work out. For me, I had nothing to lose and I could go right back to college a semester later if it didn't work. So understand your risk reward. I felt so compelled that I really had to do it and never, never looked back.

• 46:07Marc Andreessen?

  Yeah, so you're very much not doing what I'm about to describe. But a lot of people do do what I'm about to describe, which is a lot of successful people basically say, oh, you know, you should do what I did. Right? And you should like, you know, do this and that and everything that I did. And of course, that's just massive survivorship bias. You only hear from the people who left college to start a business when the businesses were hugely successful. You don't hear from the people who like the businesses crashed and burned and their life ran dramatically off course and completely fell apart.

• 46:32Speaker 03

  Yeah, I think Charles Manson also dropped out of college, but you know, so.

• 46:36Marc Andreessen?

  And much more succinct way of putting my point. Yes, exactly. And so I always get it as a little bit uncomfortable with the generalizations of how people should template their lives on extraordinarily successful people. And then I think there's another irony to it. And I think you probably agree with this. You were probably a case study of this, which is the people who are really gonna do this are not gonna listen to advice anyway.

• 46:55Speaker 03

  Bingo.

• 46:55Marc Andreessen?

  Right? So for people who give advice, this kind of thing, I'm always kind of curious as to who the advice is actually for.

• 47:01Speaker 03

  Well, yeah, I mean, if I had gone around and asked people, should I drop out of college, you know, where I'm like on a track to be a doctor and start this company to compete with IBM? Nobody would have said yes. Right? You have to be crazy. So, you know, I didn't really ask anybody, didn't really care what anybody thought. To me, it was a good idea. That was enough.

• 47:26Marc Andreessen?

  Now, the other thing, by the time you dropped out, you already had the business up and running to some extent, right?

• 47:32Speaker 03

  Yeah, I was doing about $80,000 a month in my little dorm room. And, you know, thought, well, if I had an office, I could do 10 times that.

• 47:43Marc Andreessen?

  Yeah. I think that's the other thing that people really underestimate. And, you know, your story is not, I mean, it's not a common story, but you know, Mark Zuckerberg already had Facebook running when he left Harvard. You know, Larry and Sergey already had Google running as a search engine when they left Stanford. I didn't leave, but I just, you know, I had Mosaic running by the time I started my company. Like, there's this narrative of kid drops out, starts company, and magic happens that often discounts. Like, kid actually already had thing working at that time. Right? And so maybe the actual really relevant advice is like, get something up and running, and then basically make a decision like that, which that's such a staggeringly hard thing for most people to do that most people can't do it. But that might be actually the true backstory in a lot of these things.

• 48:18Speaker 03

  Much better advice.

• 48:19Sonal?

  Is there also maybe like a certain pool of timing? Because one thing that struck me as well is that you talk about how there's kind of this frenzy that you were beginning to see when it came to the PC revolution. And frankly, I see that right now with crypto. And we have an intern on our crypto team who dropped out of Harvard. To join the crypto team full time. And part of me is like, I think she did the right thing because there is something about the times that you're in where a lot of excitement and sometimes it can be bubbly, but that community coming together is actually what innovation is. So do you have any thoughts on the pool and the draw of that kind of a moment when it comes to this decision, your decision to drop out at that time?

• 48:59Speaker 03

  No question about it. I think timing and the sort of energy of the moment are super important. For me, it was sort of the dawn of the microprocessor age, you know, IBM was going strong and it just seemed like that was too compelling to pass up.

• 49:17Sonal?

  But how did you know like that it was ready? What I think is really hard to tell, and you can only tell it in the long arc of history, is whether at the beginning or the end of something because they look the exact same. And, you know, I've been following crypto, working, observing since 2011. And so that to me is, I think, the most confusing part is how to tell. That's the thing that I think keeps people from like the FAANG companies from joining crypto, for instance. I guess that's maybe another way of asking my question. Like, what was the sign?

• 49:44Speaker 03

  It was the $80,000 a month I was doing in my dorm room, right? That was a, that was a really big sign. And I was actually still in college. And I was super excited also about the idea that anybody could have a personal computer. I mean, this was way back when it was a really new idea. Interestingly, the students weren't even buying PCs.

• 50:07Sonal?

  It's just fascinating, the whole timing thing. One other thing along those lines. So in your book, you have a scene where you talk through all the arguments that you're going to make to institutional shareholder services that your deal was one of the best ones possible. And some of the arguments, I'm going to quote for a minute. You list it in a list. Number one, this is a company I started 29 years ago and have guided ever since. It's crucial now to keep it on the right path. And I care about what's going to happen long after I'm gone. Number two, you talk about industry acceleration and then quote, transforming as quickly as possible is urgent. Change or die. Number four, transformation is rate limited for a public company. So I could keep going. There's like more. I think you have like 60 arguments there. And oh, sorry, one more quote in another chapter.

• 50:51Sonal?

  It's in a conversation with you and Egan where he says to you, quote, "This isn't really a transformation. You've already done the transformation." That was a risky part. Now is the part where you've got to grow and be on offense, which over time will create more opportunity over a longer period. So the question I have for you, and this kind of goes back to this theme of leading from public to private to public again. How do you, and this can sound very generic, but how do you manage transformation? Like not just fighting to keep your company, but to change your company? Like, what are your probably biggest lessons learned or advice for founders listening to this?

• 51:26Speaker 03

  Well, look, the pace of change in our industry is not slowing down. It's only going faster. It's the kind of industry, as you know, where like 80% or more of your revenues come from things that you just introduced this year, right? And so you have to constantly be imagining, all right, what is a successful version of our company going to be in five years? And how do we drive that change to get there? And if you're kind of stuck, you don't have the right assets, or you're trying to eke out a certain earnings per share, you can't make the investments, that's a tough place to be.

• 52:08Speaker 03

  The go-private allowed us to reignite a lot of the risk-taking and entrepreneurial spirit inside the company, certainly invest in and acquire strong capability that put us on a totally different trajectory. And we talk a lot about where's the world going with Edge and multi-cloud and 5G and all the AI and ML workloads, and how do we build out that future?

• 52:38Sonal?

  This is a good note to end on. Michael, thank you for taking the time to talk with us. I'm really glad we covered business, finance, capital, innovation. It just kind of all comes together in your new book coming out October 5th. Play Nice But Win: A CEO's Journey From Founder to Leader. Thank you so much for joining us.

• 52:59Speaker 03

  Thank you all very much. This was a lot of fun. I really appreciate it and thank you. Thank you.

• 53:04Marc Andreessen?

  Yeah, Michael, thank you for doing this. This is really great. It's really, really great to hear your voice and congratulations on the book.

• 53:09Speaker 03

  Thank you, guys.

• 53:10Sonal?

  Thanks, everyone. So you should be in the Zoom app. There should be like a preferences. Oh, my God, I'm embarrassed to ask this, but I just realized you might not actually be using a Mac, right? You're using a PC.

• 53:19Speaker 03

  No, of course not.

• 53:24Sonal?

  I don't know what to do with that.

• 53:26Speaker 03

  No, here's a trivia question for you. When was Apple's highest market share in PCs?

• 53:33Sonal?

  Oh, my gosh, maybe 1980? I don't know.

• 53:35Speaker 03

  You're pretty close. 1981, right before IBM introduced the PC.

• 53:41Speaker 03

  Okay, so you were saying?