Hyperconverged Infrastructure, VMware Disruption & the Future of Edge with Mike Matchett

Bruce Kornfeld

Welcome to PodMagic, real conversations about solving real IT challenges. I’m your host, Bruce Kornfeld, the Chief Product Officer at StorMagic. And we are always exploring how simple, reliable technology can help you and the customers you serve, whether that’s retail stores, edge sites, manufacturing facilities, or anywhere you’re supporting customers on the front line.

We always try to bring interesting guests like we have today, deliver some value and have fun along the way. So let’s dig in. I’m very excited today for our guest, Mike Matchett, who he and I have known each other for a while. We’ve had many conversations over the years, none of them video recorded, so we’re excited. Mike is the IT analyst, I guess you would say, or founding analyst at Small World Big Data. So welcome, Mike, and how about just a quick intro? That’d be great.

Mike Matchett

Yeah, Bruce. No, thank you so much for inviting me here today. I think we’ve talked a number of times over the years in a number of capacities. But, you my background going back, worked at a couple tech companies, like, for example, BMC and Dell and so forth, doing product management, marketing, and so on. Eventually went into becoming an industry analyst. I was at Tenasia Group in the days where we were covering hyperconverged infrastructure almost from the beginning. In fact, the end of the day, fact, Tenasia Group coined the term hyperconverged infrastructure. So they claim, they claim. Yeah, I know. It’s debated a little bit, but we covered that for a number of years. Of course, Tenasia Group is retired and I now run Small World Big Data and cover not just hyperconverged infrastructure, but cybersecurity, AI, and a bunch of other topics.

So I’ve become a mile wide and an inch deep, Bruce. But we get more vantage and perspective as we get older and climb the ladder and look down, right? So let’s just act a little wiser.

Bruce Kornfeld

Yeah, no, it’s OK. We’re going to have some fun today. So what I thought we’d start with is I definitely want to talk about hyperconverged infrastructure. But you reminded me of something. The word software-defined storage haven’t come out of my mouth in a while. And it might be interesting to just take a step back and think about that market. will say a quick funny story here is that StorMagic, we’ve been in the software-defined storage world for quite a while.

Our company was founded about 19 years ago. We’d been shipping product in production for about 15, and it started with Software Defined Storage. And in fact, before VMware released vSAN, we were their two-node partner with vSphere. So they had selected us as a partner. SVSAN is the name of our Software Defined Storage product. And lo and behold, a few years later, they came out with vSAN. They had slightly larger marketing and sales budgets. So they were able to out maneuver us a little bit on the market front, but we still do quite well. And customers still use our SVSAN with VMware in many, many instances because it’s simpler, easier, lower cost, et cetera, et cetera. But anyway, so we’re experts in this software defined storage space, but it doesn’t come up that much anymore. I wonder why. Do you have any thoughts on what’s happening in that market?

Mike Matchett

I think it just became, I would say, de rigor, right? It just became something that everyone became very comfortable with and part of what you expect in infrastructure. I expect to be able to load a VM or container and have that be my storage array. The days of having to necessarily buy a SAN and plug in fiber channel, long gone for most people. There’s still some use cases for it at the high end of the industry and in enterprises, but the rank and file organizations were like, hey, we’re going to cloud, we’re going to hybrid organizations, we need our storage to be more agile. And this just makes more sense to do that. probably less of a focus there.

Bruce Kornfeld

Yeah, I wonder, when you talk to end users, when you talk to clients, wonder, you probably don’t get too many inquiries about software defined storage. It’s probably all transitioned into a bundling within a hyperconverged infrastructure. It’s probably what’s happening, I would guess.

Mike Matchett

Right, think people just focus less and less on what the actual storage is under the hood, right, because more of a service over time. And I think cloud did that to pretty much a lot of different things where I want to buy a place to store data. I’m not really as concerned about managing what’s under the hood. So there’s so many different ways that that technology then bounced back and folded back. And we’ll get into some of that with HCI itself, because there’s a lot of analogies there, right? It’s just that, the software defined, it unlocked people from this mindset of having to have the silo of sands and arrays in the data center. But it’s now reverberated back and forth, as you said, in a hybrid way, such that people just say, I just really want to buy a storage service. need these data services at a certain cost and performance and so on, and less of a focus on what’s actually, is it a VM? Is it a container? What is the thing?

Bruce Kornfeld

We’ve both been around in and around the storage industry for quite a while. One positive thing about it is the quest and the need for storing bits of information never slows down. It’s amazing, which is awesome, right? It’s like for us in the storage world, it’s great because there’s always a need to store more and more and more data. Look at all these huge data centers being built by Amazon and Google and Oracle, et cetera. It’s all about storing data.

But these days, like we’ve made the transition to say, you know what, storage is important, but running applications reliably is even more important and storage is a subset of that. And that’s kind of the world that we’ve lived in now, which is we still do the SVSAN because customers are still using hypervisors like VMware and Microsoft Hyper-V and they need good software defined storage options. So we’re still in that market, but where we’ve transitioned to is a whole hyperconverged infrastructure.

I’m sure you see a lot of it. what would you say? Where do you think Hyperconverged is on its maturity curve or adoption curve? Do you think it’s still early days? Do you think we’re late in the game? What are your thoughts on the market?

Mike Matchett

I definitely, we’re definitely beyond the early days, Bruce, as you’re well aware. You said yourself, you guys were starting to do that, started the company 19 years ago and probably 15 years ago already, we’re looking at saying, how do we converge this into a smaller footprint for our customers? Our customers don’t just want software to find storage, they want software to find everything else too. And then once you do that, you start to say, can I buy this in more Lego like bricks of infrastructure and unlock, not just from the SAN and the storage thing, but how can I unlock from those big scale up monstrosities I used to do capacity planning for in the data center? know, it’s like the Spark E-systems with 50 chips in it and it’s like, know, multi-million dollar boxes and say, no, let’s look at virtual machines. Let’s look at scale out. Let’s look at the way we distribute processing and unlock that. And I think HCI is, I wouldn’t say it’s legacy yet at all, right? We are not early days, but it’s matured. And there’s definitely a lot of variation and versions out there and some very competitive offerings as you yourself can attest to. So we are, think, in a market that’s still growing, right? It’s still growing rapidly, if I understand the numbers right.

Bruce Kornfeld

Yeah, that’s exactly what I was going to say. It’s amazing to me. We run into a lot of customers. mean, our target market is the enterprise edge for sure. So big companies out of hundreds and thousands of sites. We love those customers. We love all of our customers. But there’s also a huge market for the SMB, just small businesses that need to run technology. They need to run applications on premise reliably. And believe it or not, we still run into a lot.

That have a couple of servers and they have an external SAN and they’re still doing the old fashioned way and hyperconverge hasn’t hit fully yet, I don’t think in the SMB space and it’s still evolving there. So I agree with you, it’s still, it’s not early days, but it’s certainly not fully baked throughout the industry, right?

Mike Matchett

I mean, people still have mainframes, Bruce. Let’s put it this way. The mainframe will never die, right? So there will always be SANS. There’s always these big Oracle databases or something where you need this umpteenth extra bazillion storage array. But I think you’re absolutely right. What are all the modern workloads today that we look at? And they’re all driving towards a distributed, I won’t say fully edge, but a distributed computing environment where more and more of what needs to happen is less and less in a centralized data center and more and more at the edge or office or location where business is actually happening or the data is actually being collected, the car is actually being driven, that kind of thing where HCI just has this growing market opportunity. It’s just amazing.

Bruce Kornfeld

Listen, as a vendor in this space, we have a hard job because we have to figure out who do we build for? What do we build? What’s the right priority? mean, it’s all product planning is not easy to do, right? But I’ve got an even better one, which is I wouldn’t want to be an IT director or the leader of an IT department because there’s so many different ways to solve a problem. So guess my question, I don’t know, this is a little bit off topic, but if you’re a game for the question,

Like if you were an IT leader, how do you decide what’s the right architecture? Do I run everything in the cloud? Do I run everything on-prem? How do I figure out where workloads should run? Do you have any thoughts there?

Mike Matchett

I have some opinions as a technologist, Bruce. I don’t know if I have much advice for someone who’s got to sit between the business and the technology. That’s a hard seat to be in because you’ve got people saying at a board level, we need to be all in on cloud because look at the cost opportunity here. But then they make these multi-year contracts that aren’t being used up and they’re just throwing money away. And then you’ve got other people say, hey, we should repatriate because we should have our own future and there’s all these security concerns and put our arms around stuff and bring it in and some successful examples like 37 signals and so on, right? Where they’ve made some big waves about repatriating. But, you know, I definitely think as a technologist that there is room to really optimize still where your data lives, where your computing happens. And a lot of it still comes down to using HCI in new and exciting ways because you can put that bundle of, I hate to say mini data center, but you can put that bundle of IT where you need to put it. If you look hard enough, there’s all sorts of different kinds of HCI that will fit into many different slots.

Bruce Kornfeld

Yeah, no, we see that all the time, which is typically what we hear is to put the HCI, run your application, store data as close to where the data is created. If you need high reliability, if you need low cost, cloud is great, but cloud can’t be for everything. But again, IT directors, they have a hard job to figure out where do you place on-prem? Where do you depend on cloud?

Okay. What about another, another fun topic, even though it’s been a few years now, but things seem to keep evolving. Let’s talk about what Broadcom has done to our industry and has created lots of, lots of opportunity for analysts and vendors and, and users to have churn and discussions. What, what are your thoughts on where we are a couple of years in what’s happening, you know, with VMware and Broadcom, where do you see it going?

Mike Matchett

This, mean, the easiest thing to say is Broadcom did sort of that classic big six consultancy activity where they said, hey, we’re only going to keep our most profitable, largest customers. And we’re going to, whatever we do, we don’t need to keep the rest of that pyramid behind us of clients and customers, which has pushed, I don’t know what, 70 % of the VMware customer base into looking for alternatives to VMware.

And everyone goes, what are they doing? They’re ruining it. It’s like, no, they have their own strategy in mind. was a merger and acquisition level discussion to harvest the audience that they had, right? And raise the enterprise licenses and abandon the rest, right? That’s classic, classic big six advice. So I think it just generates even more opportunity, though, for companies like StorMagic, obviously Free Stuff, we hear a lot about Proxmox, we hear a lot about a couple of other HCI vendors saying, we’ve got these great new open doors. But at the same time, it’s difficult for an enterprise vendor or even a medium to enterprise vendor who’s vested so much in VMware over the years to walk away. It’s a hard thing to do.

Even if the price goes up and even if the feature set goes down and support goes out the door, it’s just ingrained. So this is going to take a while yet. This is not something that happened overnight. It’s been a couple of years. It’s still happening. And I think we’re going to see this move away from VMware and Broadcom going for another couple of years yet as the alternatives start to step up.

Bruce Kornfeld

I definitely want to dig into that a little bit here in a second. What I was going to say about Broadcom is I agree with you. They’ve made some smart business decisions. Their revenue continues to grow. I don’t have the numbers, but I’m sure it’s pretty significant. Their stock price keeps going up. So in their minds, they’ve probably made some really good decisions here to focus on their top paying, highest margin customers. And companies like us are benefiting as well, because now they’ve opened a market for others to come in and serve the rest of their customer base that is ready for a change and can’t stay with VMware because it’s just too expensive. So it’s been a fun couple of years. We continue to see the opportunity. We’re having a lot of fun with it. Do you see, going back to that topic about leaving VMware, what do you see, what are you hearing from customers? Some are leaving now, our SvHCI product that we have, full stack HCI, we are selling it. We have customers saying goodbye to VMware and they’re joining the store magic family, but Not all of them are so what do you know? What are you saying about this migration? How long do you think it’s going to take and do you think at some point that Broadcom will stick with their top 20 30 percent and everybody else will migrate away eventually. Just give me your thoughts there.

Mike Matchett

Well, I think that’s a good five, six year outlook for Broadcom is to say, we’re going to keep our top most profitable customers, the ones we can support the easiest with the least amount of effort, and we don’t have to do a lot of R &D yet. But it will catch up to them. At some point, when you don’t do the things that VMware was doing when VMware was a healthy go-get-em-Pat-Gelsinger kind of company, the research and even the vSAN stuff they were doing, it’s going to get stale. It becomes legacy. It becomes eventually obsolete. So I think that’s on the wall for Broadcom, and they probably don’t mind because in five or six years, they’ll probably buy something else and put it behind them. This is the classic big company move. So that’s where think the market’s going go. That’s probably about the time frame it will take. And in the meantime, because the HCI market as we discussed earlier is actually growing underneath and Broadcom is not going to be in position to harvest any of that growth because they are abandoning a huge swath of their customer base and you know, StorMagic for example has a great opportunity to get that. We’re going to see a little bit of a renaissance I think in HCI. There may be some different words that people are to try to liven it up with. know, as marketing, talked to lot of marketing people, they’re going to come up with some new marketing words, Bruce.

But that’s fundamentally what it’s going to be, is just putting compute in the right place. And we’re all going to go there like, we were doing that 10 years ago with HCI, like, what’s this? But it’s, it is really chase the data, right? It’s follow the data. And so we can talk about some of the workloads coming too, that are changing, know, AI will be hard not to miss. But even other things that get pushed to the edge where time is critical, real time matters, whether it’s video or communications or just getting devices with the right information at the right time to be useful rather than something that’s a historical lookup or a dictionary, Like, we want to be aggressive with things. It’s all HCI under the hood, I think.

Bruce Kornfeld

True. was actually, I want to talk to you about, I don’t know if it’s a fun topic for you or a painful one, but it never goes away, which is backup. So I had, I had Michael Cade on who’s from Veeam an episode ago. It was a great conversation talking to him. One of the things that we run into is one of the smart things, I’m sure there was some smart people at Broadcom that thought this through is one of the reasons that leaving VMware is hard is a lot of end users have built their backup infrastructure around tight integration with VMware, and the other hypervisors don’t have that tight integration. So what we talk about is, oh, just use agents. Agent-based backup, it’s the way that backup has been done for not centuries, it’s a little bit of an exaggeration, but decades anyway.

But I mean, do you hear anything about does backup really, are people so tightly integrated with their backup solution that that’s keeping them with VMware or do you see anything around backup and change coming about?

Mike Matchett

That’s difficult question to unpack and give you the one right answer. think, yeah, I mean, there’s a couple of things I think about when I look at that data protection end of the industry in VMware and people looking for alternatives. And that is there are more and more alternatives to protecting your data every couple of months. I have lots of vendors coming in and briefing me on things. And we talk less and less actually about backup, to be honest, and more and more about data protection, data resilience, data availability, to the point again that people stop saying, what I really need to do is back data up. And then what I really need to do is recover data if it’s ransomware. What I really need to do is to saying, like, I just don’t want any of that to happen. I need 100 % immutable environment for my data so I don’t get ransomware. So we talk about cyber resilience these days. We talk about other things.

Now, again, peel back the hood and look underneath and there’s probably some, would have to be some core technologies that we go back and say, hey, we were doing this again, 20, 30 years ago with backup and backup agents and replication and sync out to the edges and so on. But the market is kind of shifting for what people want to buy and focus on to, I want to buy resilience and data resilience and data protection and things.

You just don’t just don’t let me get killed by ransomware. That’s what that’s what the IT buyer is like starting to say like and you say do backups really do that for you anymore if I just in positioning backup and saying I got a backup at the VMs and maybe not. Don’t know if backups backup the backup world strictly is keeping pace with what really has to happen right because you can make a backup of corrupted data. You can make a backup of hack. I mean there’s a dwell time for hackers that goes way back. So there’s a lot of people just rethinking what data protection means.

Bruce Kornfeld

I was going to say, it’s nice to see that there’s like a whole innovation world coming up of people figuring out new ways to protect data, new ways of handling ransomware. That’s what the industry is all about, innovation and finding new ways of doing things. We actually have a good number of customers that choose to not use any backup at their small sites anyway, at their edge sites. They have so many of them that doing a daily and incremental back-up at hundreds of sites.

Mike Matchett

How many yeah, I mean how many stories do we have of the guy at the ROBO edge who’s supposed to pull out the tape drive every night and stick and rotate the five little 8 millimeter tapes through it every night and he does that faithfully for 40 years and they figure out that the tapes have worn off all their magnetic coating and haven’t recorded anything in the last 20 years, but he’s been doing it every day, right? It’s just story after story of that that kind of thing going on.

Bruce Kornfeld

Yeah, so there’s new ways of doing it. Some don’t even do any. They send their really important data back to a data center somewhere nightly. But other than that, if they have an outers, there’s usually not enough data that’s going to bring that site down because they have some amount of replication, but they don’t need a physical backup anymore. Yeah, so there’s all sorts of ways of doing it. We don’t need to go any deeper into backup. Let’s talk about…

Mike Matchett

And again, like mainframes, tape isn’t dead either. don’t want to give you the wrong impression. Tape is actually a thriving industry too, because it turns out tape lasts a long time and it’s immutable. you know, there’s, yeah, nothing’s going away. Just to put that out there. So we’re talking about like a growing pie and what’s happening at that growing edge more than what’s happening to the old legacy view.

Bruce Kornfeld

All right, let’s talk a little bit about edge, edge size, edge computing. What are you hearing from clients and just the industry in general about physical sizes? We see there’s a lot of activity. We announced a couple of months ago an OEM agreement and a partnership with Snook, the company formerly known as Simply Nook. We have partnerships with HPE and others that have small device factors, small form factors for the edge. Are you seeing more and more of that coming along? Do think there’s a big demand for small form factor at the edge?

Mike Matchett

Oh, absolutely, absolutely, Bruce. And I think this starts to overlap with what people formerly kept separate as IoT. I think that as devices become more capable, you know, they’re not just running, they’re not just an OT thing anymore. They become a small computer. They can have the equivalent of an ARM chip in them or whatever it is, and they can run basically a full HCI stack in a smaller and smaller place. And as it does that, that definitely pushes out our ability to have distributed computing. Distributed computing at that scale has its own challenges. Don’t get me wrong. I think there’s definitely going to have to be some advances in how we know where our data is. And again, how to protect it as we were just talking about when it’s all distributed everywhere. There’s new frontiers there to explore. But as you point out, it’s probably not local backup anymore. That’s just not going to cut, you got a million devices.

I don’t care if you have a million local backups. That doesn’t solve the problem, right? We have to look at this more holistically. So yeah, and when you say edge too, I just want to throw in there. I no longer talk about edge as just its own little domain, right? It’s when I bring up edge, it’s edge to cloud or it’s edge to something. There’s a conversation that has to say the edge doesn’t live by itself. This is now a truly more distributed hybrid environment and we have to talk about how you address all those facilities, system management, security and performability and availability across that whole environment.

Bruce Kornfeld

Yeah, because if you’re a large enterprise with lots of edge sites, you’re clearly integrating edge as part of an overall IT strategy that includes data center and cloud, or one or the other at least. But even for small entity, a small company, they’re not just, if they’ve decided they need to run on-prem and have HCI locally, that’s great, but that’s not the only place they’re running applications. They also have a certain number of cloud applications as well. So you can’t avoid it. I agree with you. It’s not one or the other. It’s both and it’s multiple. you’re going back to the comment earlier, if you’re an IT director and VP of IT, you have a hard job to figure all this out, right?

Mike Matchett

And then there’s something else, another factor there that the world’s increasingly becoming a supply chain. So it’s no longer just your organization, right? You talk about Edge and Edge to Cloud and your data center and stuff, but then you look at the apps, you look at how many SaaS apps you use, you look at how many, which are just basically IT that you don’t control, right? And it’s an incredible growing amount as well. And then you look at your supply chain and say, my critical applications, my critical processes are really not within my four walls. You it’s about how that’s all coming. And so you say, we have to, we have to worry now about securing and managing that aspect as well.

Bruce Kornfeld

Speaking of supply chain, just having to think of another thing is around, this is a hardware thing, memory prices, SSD prices, with what’s going on with supply and demand, prices seem to be going through the roof. wonder, mean, how’s…

Mike Matchett

Well, who’s getting out of the business? We talked about broad, we talked about this Broadcom strategy of acquiring VMware and cutting off the non big margin parts of it, right? Well, I think you’re seeing the same thing happen with memory. And also Samsung did something with SSDs in gossip land so far, right? Where we’re not going to sell the commercial, the available stuff and the price is just going to go up. Even GPUs, like I want to go buy another GPU for my son’s AI computer. I can’t find one with the price performance that I need anymore.

Bruce Kornfeld

Yeah. And, you know, think about all these data centers being built for AI. They’re sucking up all of the GPUs and the memory, and now there’s more demand. So I just wonder, like, what do you think could happen to prices? Like, as server prices are going up, I think that’s a fact, memory prices are going up. How do you think that can impact the average IT team that has to figure out deploying servers at the edge? I wonder what impact these prices might have over the next year or two.

Mike Matchett

That’s a very good point. If you can’t get the memory and you can’t get the SSD and you can’t get the GPU to deploy at the edge, the edge is going to be stagnant a little bit, right? You can’t do it. Or you have to go to a more virtual perspective. So you start going back to saying, hey, let’s look at thin clients. Let’s look at DAS. Let’s look at the AI at the very edge, which I think is still also fostering more HCI in the data center.

As you can say, can spin up more and more resources more easily to support those kinds of workloads for the edge. So I think you can foster the edge that way, but that’s probably where I’d see some bounce back is going into virtual edges.

Bruce Kornfeld

One thing that we’re seeing along these lines is end users, I’m not sure it’s the right term, but sweating their hardware longer. Instead of replacing every three years or four, they’re doing four years or five and they’re stretching things. And that’s one of the value props, if you will, of what we do is that when a customer is running on any hardware with VMware and they want to exit to a new hypervisor. We pretty much stand alone in the industry that we say leave your hardware in place, use your existing servers for an extra year or two, take down VMware, bring SVHCI. And so then another option that we’re seeing is, live with what you have a little bit longer than you were planning. Yeah.

Mike Matchett

If anybody wants to buy a 15 year old MacBook, let me know too, because I’m sweating my hardware personally, probably a little bit longer than I should, my cars as well.

Bruce Kornfeld

Is that so you can afford Christmas gifts to your family?

Mike Matchett

Yeah, we’re going into a tighter time here. It’s definitely a thing when we start to look at what a strategy would be at different sized companies for their hardware refresh. You start to look at some of the bigger data centers and some power problems and they say refresh makes sense for us because every generation we buy, it’s a smaller footprint, uses less power, we get more performance.

But then as you move out of that that space you say people say like well, know my why do I need to replace this PC? From you know Windows 10 or Windows 11 as it starts to itself get old When all I really need is X Y or Z to run on it and give me the same capability I think I think that becomes a Real statement that we’ll see. Yeah, totally agree. I think I think we’re going to see people hang on to things a lot longer

Bruce Kornfeld

We actually, believe it or not, we get requests from customers that want to run desktop OSes as a guest OS, Windows 7? I mean, we’re talking a little bit a while ago.

Mike Matchett

Hey, there’s a lot of kiosk, there’s a lot of apps, there’s a lot of kiosk, you know, a lot of apps built for that kind of app, you know, that kind of thing that need Windows 7.

Bruce Kornfeld

And they had no reason to redesign the app, so it just runs.

Mike Matchett

It’s working fine. Why do I need to go pay a new development team to rebuild it on another OS? It’s also just going to become a legacy OS in another couple years. I want something stable that’s going forever.

Bruce Kornfeld

Microsoft doesn’t support it anymore, but I’ve never had to call Microsoft anyway, so why do I care? Right.

Mike Matchett

Yeah, and it’d probably let people look, I don’t know even know what the OS is. I’m not unlocking it. I’m not opening the box up because it’s working, right?

Bruce Kornfeld

All right, so listen, I think we’re going to close on one last question and I’m going to put some pressure on you because you are an industry analyst, so you should be ready to go with these questions. You ready? I’m looking for one, two or three, you choose how many you want to go for, predictions from you for the call it 2026, I don’t care, 27, 28, but predict some stuff for the future.

Mike Matchett

All I’m going to try to stick with HCI predictions, which is making it even harder on here. So one of the things we’re going to start to see, I think I’ve mentioned earlier how people were shifting their mindset to more of a service mindset. That’s been happening for a long time. I think HCI starts to become a service for people. And we start to see this with some vendors who saying, we will go to more consumption kind of pricing or consumption based model for HCI. Not that anybody’s necessarily there yet, but.

Even HCI as a Lego brick component will start to look like it has more of a cloud-like subscription feeling to it. because a lot of it is really internally software defined, that works out well for the vendors who are aggressively improving the software side, because then they reap every benefit under that subscription. When Amazon makes their Linux-based storage system better, or whatever their storage system is, they reap the benefit of that until they put a price breakout. I think this was going to work out with a good leverage for everybody there. So that’s one thing. We are going to see this idea of sustainability in green is morphing to one really back of just power efficiency and footprint. So there are ways where you can start to look at HCI bricks, and I call them LEGO bricks, LEGO is probably going to sue me at some point for doing that. But each modular modularity that comes with HCI is one thing, but making that modularity is more power efficient with each cycle, I think is going to be a bigger thing again.

Bruce Kornfeld

You’re not predicting HCI in space. You’re not going that far.

Mike Matchett

Well, you know, if you’re going to deploy, again, if you’re employ an internet of object thing or like a self-driving car, you’re going to have HCI in it probably. If you’re going to have things in space eventually, Mars rovers or whatever, that will historically follow that. But I don’t know if we’re going to host data centers in space for people on the ground yet in our lifetimes.

So that’s something else I think that people are calling things sustainable, but what they really mean is can you make them cheaper for us to operate and save us money and energy? Which just makes it more green, but it’s not the green that I would always be pushing for. And then let’s talk about AI and HCI. Let me even bounce this one back at you. I have my predictions on what happens with AI. What do you think happens with AI and HCI?

Bruce Kornfeld

Yeah, I mean, we’re starting to see it more and more at the edge because companies are trying to figure out how do I deploy and solve an edge problem with AI. So we’re starting to see some demand. What we see is a lot of times at the edge where we play, it’s small site, it’s two or three servers. So a lot of times they don’t need massive GPUs. So we’re starting to see some applications, some AI applications that can be run with CPU. Maybe someday it’ll head towards GPU. But the bigger picture question that I think maybe you’re getting at is, what is happening with AI? We can have some fun with this one. It is real. It’s gotten so much more advanced just in the last year. If we were having this conversation in January, it’d be different. But AI is real. It’s helping a lot. But I see a lot of effort around the industry without significant reward yet. There’s a lot of playing, there’s a lot of testing, there’s a lot of trying, but I read a lot, I’m sure you do too. I don’t hear a lot of super successful AI projects that are documented in saving money. So I just wonder if that’s going to impact the growth of AI in 2026.

Mike Matchett

I think we’re going to have a realistic pullback on the hype bubble in AI, but it is not a bubble that’s going to burst and say AI is done with and we’re just going to buy it. There’s no going backwards. didn’t, know, spreadsheets, if you look at the adoption of spreadsheets, great thing. Everyone said that we’re going to do better, faster, more productive jobs with it and everyone did, but it didn’t, you know, it’s not the thing that people sell. It’s not the differentiator anymore. It’s just a tool everyone uses.

Bruce Kornfeld

Yeah, it’s going to be fun times the next year or two. There’s a lot of different opinions. And again, going back to the IT director, if they get a business requirement that says we need to experiment with AI at this location or these locations, implementing is not that difficult as long as the hardware is available. It’s just, the results be there? And that’s the big question.

Mike Matchett

Right. So what do you also think then about, I asked this, I’m turning the tables on you, now I’m becoming the interviewer. I do this for a living. When we think about looking at AI for IT operations, AI used for better IT operations, how do you see that happening? Because that’s something that we’re seeing AI getting baked more and more into what we do. And I probably outlined at least a half dozen opportunities where something smarter than people in this episode already could contribute greatly to making things better. How do you manage at scale? How do you distribute data? How do you secure data? How do you ensure the data is not ransomware? Cyber resilience? There’s all these areas in which possibly a good use of AI within IT could arise.

Bruce Kornfeld

Yeah, I mean, if you really think about it, if you break it down to the bottom line is computers are always smarter than people. So think about the way that all of our systems are designed is that there’s a user interface. The engineering team has to think about how, if I’m a user, how do I want to access information using a keyboard and a mouse? But that’s been going on for, I don’t know, 50 years, probably more.

But should we stop, as developers, shouldn’t we stop thinking about trying to guess what information and users want from our systems and just let them get all the information they want in the way that they want to get it? And I think that’s where AI goes long-term. I don’t think anyone’s there yet, but there’s so much information, ways of controlling any kind of IT system that longer term, I see AI just fixing that completely open. So think directionally, that’s what will happen. I think it’ll take some time to get there though.

Mike Matchett

Right. And we know people are already in this broader IT industry talking about how do I create the AI, how do I enable IT to create the AI factory model of sorts, right? How do I get enable the IT director, for example, to stand projects up faster to support AI projects? He has never even heard of yet, right? These are things coming. They’re just anticipating, you know, and so we’re probably going to see a lot of that. An equivalent rise in cybersecurity because as soon as you put AI out to the edge, you’ve got data at the edge, you’ve got to secure and it just keeps getting more demanding.

Bruce Kornfeld

So basically, on the fly, you decided that you didn’t want to invite me on your podcast as a guest. You thought you’d just do it live here on ours. Is that what you’re saying? Just start asking me questions. It’s all good. It’s awesome. So thank you. Yeah. Exactly. Yeah. Mike, listen, this has been great. Thank you so much for joining. I really appreciate it.

Mike Matchett

You got to cross post. All the TikTokers, that’s their advice. It’s like show up as many places you can, cross post and get those links going.

Bruce Kornfeld

And yeah, let’s make sure we have more of these conversations over time. I’d love to have you back maybe sometime next year and just keep the conversation going. So thanks for joining. I really appreciate that. Exactly. It’s coming. It’s coming. I’m hoping to take the next couple of weeks off. That’s my goal.

Mike Matchett

Yeah, next year’s not that far away, Bruce. It’s coming fast. Yeah, and then we can do some look backs and see if our predictions were accurate and things really happened.

Bruce Kornfeld

Yeah. And thanks to everyone for listening to PodMagic, real conversations about solving real IT challenges. Before you finish, make sure you hit the like button, subscribe, send notes to all of your friends and family. Don’t forget your kids and your sisters and your uncles too. Everyone is going to love PodMagic. So thanks very much, Mike, and we’ll see everyone next time.

Mike Matchett

Thanks.