Modern IT Strategy: AI at the Edge, Broadcom, and What’s Next with Camberley Bates

Bruce: Okay, welcome to Pod Magic, real conversations about solving real IT challenges. I’m your host, Bruce Kornfeld, the Chief Product Officer at StorMagic. And we’re always exploring how simple, reliable technology can benefit you and the people you serve. Whether that’s running a retail store, branch offices, manufacturing sites, smart IT, smart cities, whatever you’ve got, that’s what it’s all about.

And you’re helping customers on the front lines. So my goal is to bring interesting guests, deliver some value, and have fun along the way.

So our guest today is Camberley Bates, whom I’ve known for quite a long time, probably in the embarrassing neighborhood of 20 or more years. We don’t look old, Camberley, but that’s just the way it goes. So Camberley, there’s a quick intro on Camberley that I’ve written here. Camberley has a storied career with leadership roles at companies like IBM, Veritas, and GE.

She’s the former vice president at the Futurum Group, having recently left in May 2025, which hopefully we’ll talk a little bit about here today, where she led research on IT infrastructure strategy, data management, and enterprise decision-making. With decades, sorry, of experience advising technology leaders and executives, she brings a grounded business-first perspective on how organizations evaluate, purchase, and deploy infrastructure today.

So welcome, Camberley.

Yeah, so I guess we can just start a little bit on a personal note about you’ve recently made a big decision to depart a role that started at Evaluator Group and then Futurum, and now you’re leaving. So, I don’t know if you want to make any comments about that. Where are you going? What are you doing? Are you just gonna be cycling around Colorado? What’s happening?

Camberley: Cycling, biking, hiking, backpacking all over the world, I’m hoping. So yes, it’s not a, it was a, we had sold the company to them, and so this was part of what was gonna happen over the long haul. And I’m keeping my toes in all of it because I’ve been doing this for too many years.

You know, I’m having one of the things I’m quite involved with is another organization called the American Society for AI. And I’m on the policy group there. It’s an all-volunteer, super smart crew. So that’s since that is moving the market so much that keeps me on my toes. That and watching whatever’s going on at Broadcom. You keep those going, and you’re done. You’ve got this all nailed down.

Bruce: It’s all good. We love Broadcom. So yeah, I anticipate we’ll be talking about Broadcom, VMware, and AI at some point in the next 15 minutes, but let’s see where we go with this. Yeah, so listen, sometimes it’s good to have a live audience so we can get questions, but I’ll try to be the one to keep the conversation flowing. But yeah, let’s just start with one around the boardroom. So I guess the question that I’d ask is, do you feel like infrastructure is making it into the boardroom? It’s a board of directors. Are they worried about infrastructure?

Camberley: Yes, and they have been for quite some time. One of the biggest issues that has been at the top of the boards is cybersecurity, which is infrastructure. The big shift on that one, the last year or so, has gone from this concern about the penetration piece of it to a bigger concern because they finally have gotten hit.

Which is business resiliency. And that’s a very, very significant concern at the board level: how do you get up there? In fact, in my past lives, we’ve done projects internally in IT that were board-reported on.

Where they were in terms of resiliency, the capability of cybersecurity, especially on the data level, because that’s where everything gets so exposed. So yeah, that’s the big, big one. The second one that came in the last two years, and depending upon the size of the company, is AI at scale. Because the cost of doing AI is so big, and they’re all going, OK, so we have to do this. We have to do this. How do we acquire the technology? Or where do we acquire the technology? Because while it is about data management, about analysis, et cetera, it’s also about this big piece of AI, of the technology that’s supported. And that also has to go into where the budget is going to come from. So it’s become board-level, C-level discussion, which I don’t think we’ve had until the last five years.

Bruce: Yeah, I was gonna ask, it just feels like it’s a difficult conversation because no offense to boards of directors, but in general, they’re not technologists, right? So do you have any thoughts, have you experienced, or do you have any advice for IT leaders that need to have a conversation with the board about infrastructure, AI, and security without getting down into the nuts and bolts of it? How do they do that?

Camberley: Well, the first thing they need to learn is how to storytell. And that is to paint it into the picture that is meaningful to an executive. Whether it is this is what would happen here, this is what doesn’t happen here, this is where the exposures tell the stories. Get the stories from your peers. Be part of a CIO or an IT director and learn those presentation-type skills.

Salespeople, really good salespeople out there, tell stories. So go call your salesperson and tell them to come in and help you craft the story. Make them good for something besides negotiating contracts. And that’s important. We hear them. I’ve got a friend of mine who teaches at Wharton, CIO, executive management, CIO. And that’s a lot of big part of what he’s doing is how do you talk to the C-level people?

So they don’t have to be the technology experts and feel embarrassed to ask the question.

Bruce: Another area, I think we all know this, that one of the key responsibilities of a board is to manage risk, right? So risk is always another one of, like you said, you tell a story and you make sure that there is a potential bad outcome if we don’t do X, right? And that usually gets people’s attention as well, right? Yeah.

Camberley: Thank you.

Yeah, yeah. And it’s, and they have been very involved with it. I mean, it’s, it’s because when you go into, when you have an attack and you have to report it, you can’t keep it quiet. Then all of a sudden it’s a PR piece. It’s an outbound, you know, it’s a Tylenol experience. Thank you very much if you want to go there.

Bruce: Yeah. I don’t know if this, I could be wrong, but this is just my, I’m just sitting here thinking about it. I really haven’t heard of pretty big, massive data breaches lately. I don’t know if tech is getting better. People are hiding it from the press better, but it just feels like maybe there’s less of it lately. That’s just my observation. Am I wrong?

Camberley: No, the hits keep coming. I think we are not as attuned to it anymore because it’s been coming, and it’s continued to increase. I mean, especially with the AI sophistication, because just as much as tech is, you know, the good guys are using it, the bad guys are using it. And so the phishing techniques, the capabilities that they have, all that stuff has just ramped up. So it’s not.

And also, what we’re seeing is that there are more prevention pieces that are in there, but that does not stop the penetration piece. I think even just today, there were, I mean, literally today, 8 billion passcodes and passwords that have been exposed. That’s a B. B. Yes.

Bruce: Eight billion? What? Wow. That’s amazing. Was that a government thing? It wasn’t a corporate one.

Camberley: No, it’s a, I was trying to read through it as really, I didn’t get through the entire thing, but they’re trying to determine where it’s coming from, what was exposed. It wasn’t exposed for very long, but it was exposed, and they shut it down pretty fast. It’s some sort of system of whatever, but if you go Google 8 billion, the day that we’re recording this, this one, when it came out. So I know this is going to go out a little bit later, but.

Bruce: Right, yeah. So you’re saying I should change all my passwords from password123? Maybe that’s not a good, secure way to protect corporate assets. Admin123, yeah, exactly. No, like, yeah, like, I’m not stupid. I’m not super smart. But like, I get some emails, and I’m like, I better click on that link. And then I stop myself. Yeah, maybe I shouldn’t. Yeah, no, they are getting better at it for sure.

Okay, so let’s talk about modernization. I think I’m trying to think of the other. Give me a second. I’ll come up with it. I can’t think of the other buzzword I was thinking of, but modernization is a buzzword, and there’s probably some reality in it. So let’s talk about what you think it means? Are, you know, IT teams should be thinking about when they think about when the board says we need to modernize. What do you think that means for IT people?

Camberley: Well, when the board is saying we need to modernize, there’s usually been a couple of things that have been targeted. One, they’re not happy with how fast DevOps is doing. So they need to modernize the development pipeline. OK, that’s one. Another one is on-prem versus off-prem. I mean, the cloud or in the end. And that’s shifted in terms of the discussion now. That modernization discussion is currently on what I should have in the cloud and what I should have on-prem.

Why would I have it in one place versus the other? And let’s rationalize this, because what they’ve also found out is that there’s a cost equation. The cloud is not cheap. It’s fast, but it’s not cheap. The third one is.

And you’ve heard the noise about this, and I brought it up already, the issue of VMware and the cost of VCF in that cost area. Part of that is saying, well, should we be on VMware? Should we move on to? Because if we’re doing DevOps, you’re probably moving to containers and Kubernetes. So do we move completely over to a Kubernetes platform or into that space? Do we stay where we’re at? Do we stay with that environment, or add another?

Layer on top of that one with Tanzu on VCF or something along those lines. Not to get too technical on that kind of stuff, items, but you know, that’s another modernization question. So I think it’s part of wherever, when somebody says we need to modernize, the next question needs to be What do you mean by that? And what do you want to accomplish at the end of the day? Am I cutting costs? Resolving my cost? Speeding development?

Bruce: Yeah. Well, listen, you brought up two topics in the last 90 seconds that both need some digging. And by the way, it’s completely unrehearsed, and actually, the topics are barely discussed ahead of time. So I’m going to throw it at you, but I know you can talk about this. So yeah, I want to talk about, before we jump into the Broadcom thing, what about the whole, you mentioned it, cloud versus on-prem. Cause that’s something we at StorMagic think about a lot.

That’s an area that we focus on, helping customers make that decision. Or once they’ve made a decision that cloud isn’t right, that’s where we come in. But I’m just wondering, what are your thoughts? How do you advise someone about what goes in the cloud? What stays local? I mean, I would almost argue that you mentioned performance, but I would have a little discussion with you about that because there are applications that need to run outside of the cloud that need

a lot lower latency than the cloud can deliver. So I just want to get your thoughts on how you advise clients on that.

Camberley: Well, the obvious ones, OK, so latency is one. I mean, if you’re manufacturing, you’re mining, you’re a retail store, all those things, you need the latency there, you know, immediate kind of capability. I mean, there are ways around some of that by doing some of these global file systems, but it’s not the same as a transactional, maybe transactional system, as more sharing of information.

The other thing that we saw this early on, I mean, I’m talking pre-COVID, where once you get to a stable environment, meaning that I know what my transaction volume is, I know what my data environment is, if you do the quick analysis, it’s always cheaper to be, almost always, I always say that, almost cheaper to be on-prem than it is in the cloud by a significant number.

And we’ve seen that, we’ve done the analysis, we’ve looked at the numbers, it’s been proven over and over again, it can be somewhere in the range of about a 30% number. So why would I do that? Well. Why wouldn’t I do that? I might not want to because I can resolve it, and I won’t have all the staff. It’s kind of like outsourcing a little bit. So I need to make my resolution in terms of where I would place it based on a cost basis in those pieces. The other thing that people like about the cloud is access to different tools and capabilities. That’s kind of the, if you will, that would be a lock-in kind of situation that you might have with an AWS or Google or Azure. But that’s similar to what we have when we have, quote unquote, talk about lock-in on-prem. It’s the same thing. You’re locked into Oracle. You’re locked into SQL Server or something like that. So those are all trade-offs. And as they get under pressure with cost pressure, they’re looking at these more diligently. And it’s not an absolute cloud first anymore. It’s a where it makes sense to place it?

Bruce: Yeah, I mean, that’s what we’re starting to hear more and more of is obviously cloud isn’t dead, but certain applications just don’t need it. And it makes more financial sense to run it yourself. But I just wonder, like, just wonder here, like, are IT people that kind of grew up in the last 10 or 15 years, are they like, maybe they don’t know how to deal with infrastructure themselves. They’re all very cloud-focused. I wonder if there’s a skill gap now. Now that the cloud has been so prevalent for the last decade or so.

Camberley: Well, there have always been skill gaps because I would rather spend my money on development and new items and that sort of thing. But what we’re seeing companies do, and I know you guys have been working on this as well, is how do we simplify the administration and the responsibilities? How do I make this so I don’t need a systems architect to run my system?

You know, can I, and so we’re using, that’s the goodness of what we’re seeing with the technologists like you, using some of the AI, their training in AI, chatbots and those kinds of things to do some of the work, to say, help me do this and, or show me how to do this. And so it will show them the entire code or the steps to do that. I automatically go through, get it done. And I don’t need a big brain on it. I need a junior or, you know, the next level, level two system admin running the system. And so that streamlines things and makes it maybe a little bit more cloud-like. I’m not sure if I would say that, but that’s what we’re probably trying to get to here.

You still need to keep within your staffing some way, whether or not it’s a consultant or whatever, your senior architects. You need to bring them in, do the analysis, and make sure you’ve done that. But it’s kind of like, I don’t need them there all the time, or if I’ve got a large enough environment, maybe I do. It just depends on the size of the company you are in.

Bruce: Yeah. Okay, I’m going to change gears here. Let’s talk about every IT person’s favorite topic in the last two years, which is Broadcom. We at StorMagic love this topic because it’s just helping our business significantly. So yeah, tell me, what are you hearing or what advice do you give clients about how to handle this juncture that’s happening in the industry right now?

Camberley: Well, Hock Tan is an extremely brilliant man. He’s a good businessman. He knows how to work with his major clients. And he had done a lot of this prediction early on, when we were rolling out that there was going to be a year to two years of decision-making about what I would do. And if I am a VMware client and the cost goes up, which it has.

We see a couple of things going on. One, the very, very large companies, some are migrating off. But a lot of them are going to stick because the cost and complexity of migrating off is huge.

Yes, if you’ve ever been through a conversion, it’s extremely painful. And with AI dropping in on you, you can’t afford to divert resources to do this. So I think that he’s lucked out a bit with the AI craze. Now, when you come to the mid-size guys, which is why I love companies like yours and that kind of stuff, because you play a lot in that market, as well as at the end.

ROBO kind of marketing that kind of said those guys he doesn’t care Hock Tan does not care as much about that market it’s not as profitable you know think about mainframe what he’s done with CA and that kind of thing so he’s not gonna bend over to make sure these guys are you know completely in game so what I look to them and what we’ve been advising them is to look seriously about what it looks like can you.

Look at your options. We’re now seeing companies move to Kubernetes or the other VM options farther along, to where they are coming up to being closer and closer to what VMware can do. A year and a half ago, that was not true. But now that we’re getting there, we’re also seeing tools coming out that are enabling customers to do some migration.

And it makes it a whole lot easier. So it becomes an option when a client gets ready to do a changeover of hardware, because you’re probably gonna have to change the hardware and everything. It becomes an option to have that kind of conversation about whether we stay or go. And if I go, am I going to a Kubernetes platform, or am I going to a VM-type platform, or what am I doing?

Should the IT organizations look at those options? What are the trade-offs on there? How fast can I move? Because I think we’ve come a long way in the two years that we’ve been around to say, okay, maybe it’s not as ugly and painful as it was when they first flipped the switch.

Bruce: Yeah, so listen, I’ve got a couple of other follow-up thoughts on that on Kubernetes, but before I go there, since it is our podcast, I’m allowed to put a little StorMagic commercial in here, which is, you are right that typically you’re changing hardware, but that is one of the things that we shine in, is that we’re helping customers make that migration from VMware vSphere hypervisor over to ours, SvHCI, but we can let them keep their server and storage infrastructure as is.

Most of the time, most of the time. It’s a good thing. But all right, so I want to talk about Kubernetes because you mentioned it a couple of times in containers. I agree with you. There is a movement happening. DevOps is using it. But do you feel that because there have been so many years of tooling and management tools in the VMware world? And as we all know, they’ve enjoyed 90 percent-ish market share. So everyone, all the IT folks, are trained on how to manage an infrastructure? It’s usually VMware tools.

So the question I have is, you’ve seen that world, it’s evolving, but are containers and Kubernetes ready for broad deployment across enterprises today, or how far away do you think that is?

Camberley: They’re implementing. They’re rolling them all out. It’s again, this is not.

When people were, when we were talking about saying you need to have a plan for this, we had to get customers and end users over to the other side. Like, this is not like installing VM, VMware, or virtual machines. This is not an infrastructure play. This is a development play. So this is more like I’m moving from the mainframe.

To open system development. So it takes a long time to take an application that’s been built on a mainframe environment and put it into that new environment.

Three-tier open system environment. And so that’s what we’re doing now. So we would have to move it from where it is today to a container-based application. And that is not just an infrastructure play. So it does take a long time, and it’s by application by application, we move over there, and we don’t do that without a concerted decision to migrate, to re-up.

Or modernize that particular application in place. Most, I think all, I can almost say, 99% of your AI applications are going onto the container and that type of platform. So we will see that move, but it could be 10, 15 years still to go. It’s just that we are going. It’s just one application at a time.

Bruce: What we’re seeing at StorMagic with our customers, because of where we focus, which is the edge of the enterprise, ROBO, and midsize enterprise, we’re seeing that there is movement to containers for sure, but there’s a mix. There’s a need for VM running applications in virtual machines that won’t disappear quickly. I kind of agree with your 10-year horizon.

But it is starting to happen now. So a mixed environment is something important, and that is something that, of course, we support.

Camberley: And I think as they look at using AI agents or whatever at the edge and retail stores and that kind of thing, you will see the need kind of increase to do the move over time.

Bruce: Yeah, yeah. OK, I think we have time for one more quick topic, and we wouldn’t have a podcast if we didn’t talk more about AI. I was like, just six months ago? I was like, I was not that I was a skeptic, but I just felt like it was being over-talked about. But I don’t think it is anymore. It’s just I don’t know if you feel it, but the last time it was improving.

Every month, every week, it just seems like models are getting better, and they’re being used more frequently. It’s just accelerating. I guess my question is, obviously, there are personal non-business uses for AI. We can talk about that on a different podcast. But yeah, in the IT world, what do you think is either happening right now, or where do you think IT teams should start looking? Because it’s new for everybody. So where should the average IT director, VP of IT, where should they be thinking about putting AI to use for them?

Camberley: Well, the two obvious ones for IT use are DevOps. And we are seeing some of the coding work go that way, or doing tests and quality control, and that kind of stuff on code. The second area is that they should all be asking their vendors, Where are you using AI to streamline and simplify my world? OK, so we’re done with that one. So let’s now go to the bigger ones, which are the, let’s say, since you guys deal with the retail folks, how are the people that run retail operations looking at using AI?

And what are they going to be doing with it? So they’re looking at all kinds of different ways for suggestive marketing. I would say they’re looking to optimize warehouses. They’re looking at, if you look at what Amazon is doing with their robotics and sending out robots into their entire warehouse and eliminating people having to walk 10 miles a day around their warehouse. That’s a huge number.

You know, this, the gamut of applications that are coming out, they’re using this technology, is pretty significant in terms of changing how we are doing things. And I think that we want to look at it as an augmentation of what we do as business people.

Is that we can take whatever we’re doing and add on some additional knowledge to our decision making, as in managing the warehouse kind of things, etc, or what we’re doing is taking away things that are laborious like walking 10 miles in a warehouse every you know for picking purposes so each every company I think is going through that what’s great is what’s happening now is this next generation of agentic kind of things and APIs it’s going to start streamlining to the point that the companies are not going to have to do all of the development themselves and we’re going to be seeing companies come out with technologies that they can just implement, much like an application, as opposed to developing from scratch.

Bruce: Yeah, one thing that we see, again, because of where we focus, there is concern out there that there’s a perception that to run, to get the best answers from AI, you need expensive kit, right? Whether that’s cloud-based or your own, it’s the whole GPU conversation, and NVIDIA is crushing it just like Broadcom is, right? But…

There’s a movement for smaller sites. I think there’s a perception out there that it’s unreachable technology to run your models locally, run AI locally, but you don’t necessarily need expensive NVIDIA GPUs everywhere you go because there are smaller sites. Exactly. So you can use CPUs. It might be a little bit slower, but getting an answer in 45 seconds or a minute might be okay, versus waiting for the cloud and doing modeling and expensive stuff that way. So there’s an evolution happening, I think, on the infrastructure side of AI as well.

Camberley: Well, I think what you have to separate is the training from the inferencing, right? Inferencing doesn’t take a whole lot of GPUs and some inferences can be, and I’ve seen it done. We’ve tested it. We’ve proven it can be done on just CPUs. So we’re not talking about the inference, the training, which is taking whatever foundational LLMs or other LLMs or maybe some automatically developed by somebody, maybe from a SAP or some of the things that Oracle is developing or Microsoft is developing. You take that and then you do some additional training with your data, and that becomes an … Yes, you’ve got to go back and keep on training it and updating it, but that doesn’t mean I’m doing that all at the edge constantly. So there is a data pipeline that you need to understand about where it’s going to be sucking up all the CPU power and the energy, as opposed to what we’re doing when we’re executing at the edge.

Bruce: I completely agree. All right, Camberley, I think we went a little bit longer than we were planning, but it’s always hard to stop. It’s all good. So thank you very much for joining. I hope that we can entice you to get off your bicycle at some point, maybe a few months down the road, and come back on. So it’s always a pleasure. So thank you so much. And for those of you out there listening, thanks for tuning into PodMagic. Where We’re simple, reliable IT meets real-world impact. And if you enjoyed what you heard today, we’d love for you to subscribe and share. Until next time, thank you.

Camberley: Thanks for the invite.