Transcript

[00:00:00.170] Ned: Like this could be so simple if you just use a normal operating system and browser.

[00:00:06.450] Chris: I well, would prefer it if the only people that were spying on me for no good reason were my neighbors and not Google and Microsoft.

[00:00:16.370] Ned: Your neighbors have plenty of good. Call me old fashioned. Old fashioned? Do you want me to make you an old fashioned? Just call you old fashioned?

[00:00:27.610] Chris: Well, it is a Monday, so any little bit of help I can get.

[00:00:33.850] Ned: It’S a literative, too. I’m like a muddled Monday. An old fashioned would be perfect, we think. Better than a manic Monday. So sorry. I stole that right out. Yes.

[00:00:48.750] Chris: It’s like a cartoon where you pull the rug out from under me and.

[00:00:53.070] Ned: There’S actually a hole there, but you levitate for, like, a solid 3 seconds before you look down and realize there’s a hole, and that’s when you fall.

[00:01:00.400] Chris: Fall and disappear into a puff of smoke. Yes.

[00:01:03.890] Ned: So I wonder, you tell me what you think about this. Do you think if the cartoon character never looks down, they would just continue to levitate forever?

[00:01:13.430] Chris: I do believe that those are the rules, yes.

[00:01:15.690] Ned: Okay, I just wanted to check. I’m not up on my cartoon physics, but that seems important to know.

[00:01:23.530] Chris: Did you even go to college?

[00:01:26.010] Ned: Clown college. That’s a whole different branch of physics if you didn’t already know clown physics. Yeah, more or less. Yeah. I watched Who Framed Roger Rabbit recently, apropos of nothing, and it stands up. It’s a great movie.

[00:01:47.710] Chris: I was going to say.

[00:01:48.470] Ned: Yeah.

[00:01:48.630] Chris: It remains excellent.

[00:01:51.890] Ned: Okay, that’s all I need to say. Hello, alleged human, and welcome to the Chaos Lever podcast. My name is Ned, and I’m definitely not a robot. I am a real human person who is not using virtualized human organs to synthesize the appearance of organic life. That would be strange. Strange and beautiful. I mean, awful. Yes, awful. With me is Chris, who is also virtually here. Hi, Chris.

[00:02:22.350] Chris: That would also be the creepy aliens at the end of the movie AI. Which, if you think about it, were also way too unnecessarily narrow and fragile. And also that movie was dumb.

[00:02:37.650] Ned: All right, counterpoint. I’ve never seen AI. Wait, was that the one with the boy who becomes a real boy and then he dies at the end and his parents like something I vaguely remember something and didn’t have Haley Joel Osmond in it.

[00:02:57.380] Chris: It did.

[00:02:58.440] Ned: Okay, so I have seen this movie, but I seem to have wiped one detail correct. You wanted to be a real boy, right? It was like pinocchio.

[00:03:06.980] Chris: I wanted to be a real boy. Yes.

[00:03:09.710] Ned: But he wasn’t.

[00:03:10.570] Chris: Then there was all of the rest of the movie.

[00:03:14.590] Ned: I think that’s where it lost me. The rest of the movie?

[00:03:18.790] Chris: The movie that you didn’t see lost you?

[00:03:21.650] Ned: Yes. Well, there’s a reason why I don’t remember it, I’m sure, and it’s maybe because I wiped it from my brain. Maybe it’s because I want to be able to watch it again. Or maybe it’s a Men in Black type situation. It’s really hard to say. Another movie that stands up. Are you?

[00:03:39.900] Chris: Men in black?

[00:03:40.920] Ned: Yeah, men in Black totally holds up.

[00:03:43.910] Chris: Yes, definitely.

[00:03:45.510] Ned: Okay. I’m glad we’re in agreement with that, because if we weren’t, most importantly, they.

[00:03:50.270] Chris: Never made any sequels.

[00:03:53.930] Ned: None that I’ve actually seen. And that’s true. Except for Wild Wild West, which is more of a prequel when you think about it. I want that to be true. Let’s just pretend it’s true.

[00:04:09.550] Chris: I still contend that that movie is so bad, it’s good.

[00:04:13.070] Ned: I don’t think anybody would disagree with you on that. And of course, we’ve all seen the Kevin Smith talk where he talks about the giant spiderman in the Superman movie. Yeah. Okay. All right. So maybe maybe we’ll link that as the special extra bit in the newsletter.

[00:04:31.350] Chris: It’s just amazing story. Amazing story.

[00:04:34.970] Ned: True. Who knows? But definitely worthwhile. That’s what matters. So let’s talk about VMware, shall we?

[00:04:46.570] Chris: Might as well. It is on the schedule anyway.

[00:04:50.990] Ned: Well, you said I was in charge of a main, and I didn’t have anything until this morning. So we’re going to talk about VMware Explorer because I was looking through my podcast feed, and there was a Gestalt it roundtable podcast that discussed whether or not VMware should just stick to being a hypervisor and, like, Vcentery stuff or if branching out was a good idea for them. And so that got me thinking about Explore, and I did a bunch of reading, and here we are. All right, last week was VMware Explore 2023 US, formerly known as VMworld. And probably in everyone’s brain, it will always be VM world. Kind of like how Twitter will always be Twitter, right? I refuse to acknowledge anything different. The conference was renamed in 2022. I e last year. For reasons I have no idea, they decided it wasn’t the whole world for those of a certain age. Again, me. VMware was a pivotal, no pun intended technology that transformed our use of compute, storage, and networking in the data center. Now, Chris, you and I talked a bit about VMware, the company, when Broadcom announced their acquisition, but I thought it’d be interesting to revisit the company and pontificate on the future.

[00:06:18.860] Ned: Do you recall that episode?

[00:06:22.010] Chris: Vaguely, but I definitely recall enjoying Pontification.

[00:06:26.250] Ned: Yes, and don’t we always? I think that’s true of all podcasters, isn’t it?

[00:06:31.740] Chris: Now, if you’ll excuse me for just 1 second, I’m going to go put on my largest hat. Okay.

[00:06:37.860] Ned: I hope it’s a miter.

[00:06:40.350] Chris: Goes without saying.

[00:06:41.780] Ned: Okay, just making sure. So, as a quick bit of history, VMware was founded way back in 1998 when I was in my second year of college. We were so young then. Anyway, they came out with their first real product in the early 2000s, VMware Workstation. And that was a client side product meant to help people run virtual machines on their local desktop. Shortly thereafter, they launched a server class product named GSX, which was then renamed ESX, which then eventually became the stripped down version called ESXi, because they decided to make it more difficult to say.

[00:07:22.990] Chris: I was sure at the time, and I remain sure now, that they did that to make fun of Apple, put.

[00:07:31.090] Ned: The I at the end just to mess with it’s. Certainly I can’t think of a reason why they put the little I at the end.

[00:07:41.890] Chris: I mean, it’s supposed to be, I think, integrated because it couldn’t be embedded, which is what it really mean. The reason the ESX versus ESXi is as simple as the operating system and the hypervisor are the same thing. And that’s also why the installation is so absurdly small.

[00:07:59.800] Ned: Right. Because they stripped out all the components in ESX. They gave you the service console and a shell that you could get into and do things they’re like. Nope, you don’t get any of that. It’s just a hypervisor. That’s it.

[00:08:15.150] Chris: Yes.

[00:08:15.940] Ned: Which was a net good thing from a security perspective, and also ease of installation.

[00:08:21.010] Chris: Right.

[00:08:22.610] Ned: So VMware helped us usher in the X 86 era for server compute, and that change was kind of already underway, led by Microsoft and its Windows Nt server operating system, which provided a viable alternative to the very pricey Unix variants from vendors that required you to buy their hardware as well. Now, those vendors lingered around for quite a long time.

[00:08:50.810] Chris: They’re still around.

[00:08:52.220] Ned: They’re still around, and so is their Unix variants. I believe you worked on a project to help migrate some folks off of HPUX that was running on a superdome onto Red Hat Enterprise linux running on VMware. Right.

[00:09:09.390] Chris: That was one I’ve done something similar for as 400, and I still have to look around every dark corner to make sure that IBM is not following me.

[00:09:17.730] Ned: Well, they own Red Hat now, so they’re probably less angry. Yeah, so, I mean, that’s how things were. You’d have to buy the hardware and the operating system from the same vendor and they could charge you however much they wanted, and that was fun. And pay for support and support forever. And then commodity X 86 servers and VMware and Windows all lowered the cost of entry for companies that needed to process data and host applications. VMware further lowered the cost of entry by abstracting the hardware, increasing utilization of that hardware, and enabling workload portability. The launch of Vcenter to manage multiple hosts and their proprietary Vmotion feature to seamlessly move virtual machines without powering them off was nothing short of revolutionary in the data center. And I can’t stress that enough. Vmotion was a goddamn miracle.

[00:10:18.870] Chris: It was such a change in operating procedures that when it first came out, a number of allegedly intelligent CIOs banned the practice.

[00:10:31.450] Ned: I remember this. Yeah.

[00:10:35.210] Chris: Surely such a thing is impossible.

[00:10:38.270] Ned: Not only impossible, but highly disruptive. Things were going to break. And as someone who has vMotioned thousands of virtual machines, I can say even when it didn’t work, the workload still kept running on the source machine. It was pretty foolproof. And that really more than anything else, I think is what helped solidify VMware as the leader in enterprise virtualization.

[00:11:07.990] Chris: Yeah, and I’m sure you remember Microsoft running some pretty delicious FUD campaigns about if you tried to Vmotion SQL, it.

[00:11:17.070] Ned: Would crash, which was demonstrably not true. Right. You know, we can test that. Microsoft right. It’s not hard, you just right click.

[00:11:32.740] Chris: And it’s right there in the options.

[00:11:35.310] Ned: And then suddenly they introduced live migration for HyperV and everything was okay again.

[00:11:42.370] Chris: That was an oopsie. It’s all actually fine.

[00:11:46.120] Ned: Yeah, it’s all fine now. So Vcenter and sort of Vmotion and all that jazz came out around 2004. And that represented a time of incredible innovation and really the beginning of VMware as an enterprise behemoth or what it would eventually become. It was also coincidentally the first year of VM World in San Diego, California, with a meager 1500 people attending.

[00:12:15.370] Chris: I wasn’t at that one, but I was at some not that far off. And I absolutely remember it being a four digit conference.

[00:12:24.430] Ned: Yeah, well, it definitely got into the five digits pretty quickly.

[00:12:31.170] Chris: Yeah.

[00:12:33.170] Ned: So that brings us to VMworld. And now VMware explorer. One of the things that VMware did was allow heterogeneous compute environments to flourish instead of buying your hardware, your software, and your services, all from a single vendor, also known as the single throat to choke. Isn’t that fun? You could combine a whole bunch of disparate vendors in a way that actually worked for you and not for the vendors and VMware.

[00:13:02.400] Chris: Just by that we of course mean your finance department. Jim Dell servers are cheaper this month.

[00:13:13.870] Ned: Well, dude, it’s a dell. Yeah, exactly. So it led to the ability to have a hodgepodge of hardware in your data center and it didn’t matter. And the fact that you could just seamlessly migrate your workloads from one set of hardware to another made it that much easier to retire old hardware, or spin up on new hardware, or even test new hardware that you weren’t sure about. In a very similar manner, the expo floor at VMworld was a proliferation of diverse vendors, all brought together under the big umbrella of VMware servers, storage, networking, security monitoring, et cetera, you name it. All types of vendors were welcome to join, mingle strategize with each other, big melting pot kind of situation. And you also had new classes of vendors arising in the ecosystem. You had companies like Veeam who could not exist prior to VMware because they specialized in protecting data on VMware. At first. I know they’ve kind know, branched out since then, but originally they were just VMware. And in addition to that, virtualization hosts now had to be monitored, secured, patched, which brought on a whole new host of companies. So VMworld was a chance for these companies to be noticed by investors and users alike.

[00:14:43.370] Ned: Speaking of users, VMware created a new class of sysadmin called different things. Usually VM admin or virtual admin or something along those lines. And they were different in the sense that they really had to be a bit of a jack of all trades.

[00:15:00.130] Chris: What’d you call me?

[00:15:01.680] Ned: Yes, you heard me, John. So, to be a proper VM admin, you needed to understand the intricacies of VMware itself, but also you needed to understand things like storage, networking, server configuration, maybe a little bit about databases, and also identity and access. You kind of need to know everything. And if you were working with Vcenter and its available APIs, many VM admins began to automate processes using PowerShell or the VMware CLI. It was an interesting time.

[00:15:39.550] Chris: That was one of the things that you were able to do in terms of deploying ESXi and scripting all of the configuration well, before VMware had baked that functionality into its platform.

[00:15:52.850] Ned: Yeah, long before host profiles, plenty of people were fully scripting out the installation of ESX.

[00:16:01.330] Chris: And none of us commented any of the code.

[00:16:06.230] Ned: Why would you? That’s job security, baby. Try to figure out what this line of code does. So while larger organizations had silos for server storage, networking, backup, security, et cetera, VM admins sort of stood astride of all these silos, needing to interact with each one to keep their virtual estate in good order. And so, as a consequence, VMware user groups started sprouting up across the world as the use of VMware grew and the virtualization admins, they felt a need to share knowledge, build a community, drink, know, the usual stuff. The culmination of this community was a pilgrimage to VMworld, which had now expanded to host events both in the US. And EMEA, depending on what was closest for you. So attendance of VMworld topped out at about 23,000 people in 2018 for the US version. Now, while still incredibly relevant for VM admins and vendors, the rise of the public cloud and a hint of kubernetes in the distance had diminished. The once shining star of all tech conferences, AWS Reinvent, had taken VMworld’s place as the one tech mega conference that you had to attend if you were anyone, which is why I was never there.

[00:17:38.210] Ned: Last year, AWS Reinvent had an estimated 51,000 people, and that’s pretty high, considering it was post COVID and if the Internet’s to be believed, a dubious proposition. I know the last pre COVID one in 2019 had over 60,000 people at Reinvent. That’s more than double what VMworld ever had.

[00:18:05.850] Chris: The cloud’s big.

[00:18:09.390] Ned: So I’ve heard. So, with all that in mind, it’s interesting that they chose to rename VMworld. I don’t know if that was to sort of shift the attention a little bit, but the question still remains. Is VMware Explorer, and VMware, by extension, still relevant to the audience and the market? What do you think?

[00:18:37.590] Chris: I’m going to go with and I’m going out on a limb here.

[00:18:41.180] Ned: Whoa.

[00:18:42.470] Chris: Maybe.

[00:18:44.230] Ned: How dare you, sir. Now, I would say that, broadly speaking, the answer is yes. VMware still happens to be a highly profitable company with healthy margins. Enough so that Broadcom set their eye on them in May of 2022 and offered up $61 billion in cash in stock. You don’t do that for a stinker of a company. Well, unless you’re WeWork.

[00:19:13.890] Chris: Considering ESXi was and remains the bedrock of the vast majority of on prem server and data center solutions. That’s not something that people are just going to replace overnight, right? Even if they wanted to, despite all.

[00:19:33.450] Ned: The noise about public cloud, it still only represents 10% of total It spending. VMware is a decent chunk of that other 90% of spending. So not a bad purchase. And while regulators around the globe have taken a bit of convincing that Hawk Tan, the CEO of Broadcom, is not just going to fleece customers and destroy competition, the approval from the UK was bestowed last week, and it appears that the deal will actually close at the end of October. Tan has indicated that he will be investing $2 billion in VMware once the acquisition closes, with one half dedicated to R and D. That’s a nice chunk of change. How is VMware positioning itself now that it appears that the takeover is imminent? And what are they likely to focus on as a new Broadcom subsidiary? Well, we can take a look at what they presented during the keynote that happened at VMware Explorer. That should be a pretty solid indication of what they’re focusing on. And sure enough, there’s two key areas that VMware is focused on. Do you want to try to guess one? It starts with A, ends with I alchemy. I guess if you spelled it in the right language, maybe it’s got about as much nonsense in it, but no, it’s AI.

[00:21:11.570] Ned: Private AI.

[00:21:13.030] Chris: Yeah, I was totally going to guess.

[00:21:14.550] Ned: That, because of course and the second one is multi cloud, so let’s dig into those a little more. AI. Have you even heard of it, bro?

[00:21:27.610] Chris: Even? Do you even AI?

[00:21:31.230] Ned: I do not, sir. I respect myself. Well, no, you don’t. If there’s one technology that has absolutely dominated the tech headlines for the past year, it’s artificial intelligence. Now, I know it’s bad when my casual acquaintances I see on vacation strike up a conversation with me about AI. I’ve done my absolute best not to throttle them, Chris. My apologies to Derek. I hope he has a swift recovery, but it is just everywhere unavoidable. Even so, why wouldn’t it be part of VMware explore, right? Yeah. So VMware has announced a private AI product partnership. Well, it sure is something. And they’re doing that something with Nvidia, who is fucking crushing everything right now with a $1.15 trillion market cap.

[00:22:37.170] Chris: That’s just an absurdity.

[00:22:39.730] Ned: I really wish I’d held onto that stock from 2007. Not bitter. Not bitter at all.

[00:22:47.370] Chris: Chris no, you sound super not bitter.

[00:22:51.930] Ned: Yeah, I don’t want to go off on a tangent here, but I sold some stock to buy the house that we’re currently in, or at least for the down payment for the house. And among the stock that I sold was Google, Amazon, Apple, and Nvidia.

[00:23:08.090] Chris: You’re not good at stock.

[00:23:10.510] Ned: I mean, don’t get me wrong, all of those stocks had already made me some money, so I wasn’t exactly mad at them. But if I had known, I would have found the down payment another way. Chris anyhow, essentially, VMware and Nvidia have worked closely together to deliver a turnkey solution for private AI development. Instead of renting compute from one of the cloud providers or using a publicly available LLM, you can train your own in house LLM on your own hardware. As we’ve covered data leaks from public AI LLMs are not exactly unheard of Samsung, so there’s a fair number of large enterprises that would very much like to do their AI inside their own walls, and there’s a lot of money in that. So VMware, for their part, can virtualize the CPU and GPU leveraging graphic cards from Nvidia to provide the raw horsepower that’s required for AI development. Now, I have to admit that I’m not sure how useful virtualization actually is in AI, where, as far as I understand it, workloads are meant to run in memory and make maximal use of all the hardwares. So I don’t know, I kind of assumed a lot of it was running on bare metal, but apparently VMs can still play a part.

[00:24:43.210] Ned: I don’t know.

[00:24:45.210] Chris: I mean, that’s a scheduling question, right? If you schedule things properly, then the overload or the amount of god damn it, I don’t remember any words right now.

[00:24:58.750] Ned: Cache coherence.

[00:25:02.450] Chris: Dictionary.

[00:25:04.450] Ned: I don’t think that’s what you’re looking for.

[00:25:06.850] Chris: The difference between running bare metal CPU GPU and virtualized CPU GPU, assuming that it is scheduled properly, is really not that big. So then you get to the same issues that we have years and years ago, where you’re sacrificing let’s just call it 5% of overhead.

[00:25:27.780] Ned: Sure.

[00:25:28.680] Chris: Or the ability to swap workloads in and out as needed, which means you don’t have to buy redundant hardware. Will this be appropriate for absolute bleeding edge? The fastest of the fast, the most important of the most important workloads? Unlikely. But those are going to be edge cases anyway. What we’re talking about here is commoditization, which means even gigantic companies running at 95% efficiency and saving themselves 80% of the purchase price year over year, they’re going to take that.

[00:26:03.350] Ned: Yeah, I suspect you’re right. So maybe there is a part for VMware to play. And they certainly think mean AI is a growing sector and they want to take advantage of that. So I think it’s smart for VMware to get involved with the industry leader that is Nvidia. And most private data centers are already running VMware anyway, making VM admin skills transferable to this hot new area of AI. So I don’t know if it’s going to move the needle in any large way in terms of VMware sales, but it’s probably a nice addition to what they’re already doing.

[00:26:44.530] Chris: Yeah, my guess is that’s one of the reasons that the partnership is so important.

[00:26:49.170] Ned: Yes.

[00:26:49.760] Chris: Because you get the partner to say, we certify or we promise, or we’ve proven that this works on VMware at such and such a rate. That’s very different than VMware saying it.

[00:27:02.790] Ned: That’s true. And they have jointly come up with a reference architecture that hardware vendors can then test themselves against and verify. So if you’re an ISV or a vendor, you can validate your design based off their reference architecture. That could help out some hardware vendors, too. It’s all one big community, right, Kumbaya?

[00:27:29.510] Chris: We’re all friends here.

[00:27:31.140] Ned: Sure.

[00:27:33.890] Chris: Everybody brought their ceremonial conference swords?

[00:27:36.590] Ned: Yes.

[00:27:37.040] Chris: Okay, let’s begin collaborating.

[00:27:41.090] Ned: The melee begins at eight. So the other big one was Multi cloud. Now, as I mentioned earlier, VMware’s big thing was the ability to unify heterogeneous hardware environments and create, let’s call it a common substrate for people to build on.

[00:28:00.570] Chris: Let’s call it that.

[00:28:01.720] Ned: Yeah. Your ISVs out there could provide a virtual appliance image to distribute their software. VM admins could build their virtual machines across all their data centers, regardless of what hardware or storage was lurking underneath. By virtualizing and abstracting the networking, storage, and compute layers, VMware became the lingua franca of modern data centers. And then the public clouds rolled in and fucked it all up. Seriously, we had a real fine. Everything was good. We’d finally established what was basically a standard proprietary, sure, but at least it was a standard. Then Public Cloud came along, and sure, they use VMs just like VMware, but none of the major CSPs used the ESXi Hypervisor or Vcenter or the same hypervisors across each other. AWS started with Zen and then built their own thing. Microsoft made Hyper V a real boy, and Google went the KVM route because of course they did. Now, this probably has something to do with the massive licensing costs of VMware. I’m sure none of the CSPs wanted to pay that, and they’re like, oh, free, or we already own this, and that’s what they did. But the practical upshot is all the big CSPs run their own proprietary everything below the operating system.

[00:29:36.080] Ned: It’s different networking, it’s different storage, it’s different compute, none of it’s compatible. It was as if we took a giant step back for interoperability. And that’s because the name of the game in Cloud is Capture and Keep. If you can get a customer to put their workload on your platform, then make it as annoying as possible to leave.

[00:30:01.630] Chris: Yeah, I mean, it’s a fair point because a lot of the times people would think, what a virtual machine? I can run that anywhere.

[00:30:08.380] Ned: Well, no, you might think that, sir, but you would be wrong.

[00:30:15.190] Chris: I mean, even something as simple as migrating a virtual machine from Hyper V to Azure, considering Azure’s backend is ostensibly just Hyper V. Well, it isn’t. It’s Azure’s Hyper V. Yes, and while in many cases it works fine, there’s.

[00:30:32.020] Ned: Still a checklist, there’s still enough difference that it matters. Yeah. Now, the practical upshot is that VMware would very much like to step in and become the lingua franca of the modern cloud era as well. They already have their VMware cloud hosted and managed offerings in all the public clouds, which they’ve talked about ad nauseam. I’m not going to rehash that. But they’re basically the same as the self hosted option, but through products like Tanzu and I think it’s called Cloud Controller or something, they’re trying to extend into the CSPs for real and manage those workloads. Running within VMware’s foray into Kubernetes has morphed into what they call Tanzu, which is ironically not an invasive vine, but sure seems to act like one growing over and covering all of their other products in the software suite. They really should have called it Kudzu if they were being honest. Yes, I feel good about that.

[00:31:40.930] Chris: I know you do. That’s why I’m giving you a pity laugh. I mean, real laugh.

[00:31:45.240] Ned: A real laugh, because I’m a real boy. Tanzu is now integrating aspects of Aria which if you’re not familiar with that, is what was Vrealize? That whole family of Vrei’s automation orchestration monitoring. They renamed that whole thing to Aria last year because reasons and now it’s being slowly taken over by Tanzu. But the idea here is to add automation and management of AKS, EKS and GKE, among other things. And it’s also tacking on a platform engineering portal. OOH, you can buy platform engineering, a developer portal and even cloud health to do FinOps monitoring. Truly a tech buzzword bingo card packed to the gills. It does everything.

[00:32:41.640] Chris: Wait, I thought it was a vine. Now it’s a fish.

[00:32:44.310] Ned: It’s a vine fish. A fish vine.

[00:32:47.180] Chris: An underwater vine.

[00:32:49.350] Ned: Kelp. It’s kelp.

[00:32:50.960] Chris: I think a fish vine is called an octopus.

[00:32:54.150] Ned: A kelpus. We’ll workshop it.

[00:32:58.330] Chris: It’s a superfood.

[00:33:00.810] Ned: So aside from the multicloud thing and the private AI thing, there were some other things that they announced at Explore, including their intentions to extend to the Edge and build out their workspace one product more. Now I think that the Edge stuff has a chance. I’d say the workspace aspirations are a little more in vain and that’s because they are about to have new masters who aren’t exactly known for their patience and willingness to expand unprofitable market segments. I’ll also say that markedly absent was anything about their security portfolio and products like carbon black. That could be a sign that security is not really a product and more of a feature, or that they’re planning on selling those segments off as soon as the acquisition completes. I’d go with the latter, likely. So, getting back to Broadcom and how it impacts this whole scenario, there’s some really good analysis on silicon angle SuperCloud notwithstanding, regarding how VMware and Broadcom break down financially, as of the most recent annual results, broadcom did 34 billion in revenue and VMware did 13.4. That’s pretty good. I mean, Broadcom is definitely getting a decent engine there. What’s important to understand is the difference in expectations for a couple other metrics, most notably revenue per employee and their EBITDA margin.

[00:34:38.970] Ned: Broadcom averages $1.7 million in revenue per employee, and VMware is at 350,000. EBITDA margin for Broadcom is 58% and VMware is 25. Those numbers are smaller, in case you missed it.

[00:35:00.110] Chris: Yeah, it’s been a while since math class, but I worked that one out on my own. Thanks.

[00:35:04.790] Ned: Okay, so the point is that those numbers, that Broadcom, they have 20,000 people, and they make more than double what VMware does. At 38,000 employees, VMware is about to shrink, and layoffs are going to be severe. A 50% cut in headcount is definitely not out of the question. Now, some of that will involve consolidation of redundancies. You don’t need two HR departments. You don’t need two I don’t know what’s a useless department. Finance. You don’t need two finance departments. So sorry. Well, no one from Finance listens to this show. It’ll be fine. But it seems safe to say that all nonprofitable areas of the business are going to get a serious haircut. So if your group wasn’t directly mentioned in the keynote, I would not consider myself.

[00:36:04.950] Chris: Yeah.

[00:36:07.350] Ned: Part of the reason Broadcom can pull in such high revenue on such low headcount is the fact that they target mature technologies that don’t require much in the way of marketing and sales or research and development. They buy technologies that have customers who are, let’s say, locked into the tech, and Broadcom just keeps things on life support as long as they reasonably can. I’m not saying it’s rent extraction, but it’s not far off. With VMware being acquired, are they going to follow the same path as previous acquisitions, like CA, for instance? Or will Hoctan dump the promised R D dollars into helping VMware innovate? Once again, based on the analysis coming out of Explorer, it certainly seems like they’re going to focus on core competencies probably a good idea, and make a run at this multi cloud relevancy thing. I’m sure Haktan is going to give them some leeway to prove the effort is worth it, but probably not too much. There’s going to be a culling around Christmas time, and I hope the folks at VMware are suitably prepared.

[00:37:20.910] Chris: Yeah, I always enjoy when you end things on such a high note.

[00:37:28.110] Ned: I mean, don’t think of it as a layoff. Think of it as. An opportunity.

[00:37:34.050] Chris: What I heard was don’t think of.

[00:37:37.970] Ned: Oh, yeah, I do know more than a few people that work at VMware, and I certainly hope that they land on their feet regardless of what happens. So hearts go out to the employees. I think that you’re making plenty of money, and no one needs to get laid off, but it’s really not up to me. Hey, thanks for listening or something. I guess you found it worthwhile enough if you made it all the way to the end. So congratulations to you, friend. You accomplished something today. So now you can go bask in the glory of a midnight sun, whatever the hell that is. You’ve earned it. You can find more about the show by visiting our LinkedIn page. Just search chaos. Lever, or you can go to our website, chaoslever. Cow. I mean, where you’ll find show notes, blog posts, and general tom foolery. We’ll be back next week to see what fresh hell is upon us. Ta ta for now. You know, I just left cow.

[00:38:35.250] Chris: Yep. I am now furiously googling to see whether or not cow is a possibility.

[00:38:41.320] Ned: It is not. I checked, and I’m furious about that. We should really petition whoever I can to get it added to the TLDs.