The Savvy CIO: Episode 6 – “The Renew, Refresh, Replace Dilemma”

Rob Brothers: The aging technology still has to be there, right? Because generally, when we’re talking about new initiatives, running your existing infrastructure on the backend, so you want to make sure that you have somebody that you can turn to that can definitely run that infrastructure. If you think about, ” Okay, as a CIO, where should I be going?” You’re probably looking at bringing in somebody who understands the processes within your company to be able to say, ” Hey, what should I be doing here? What should I consolidate down?”

Bradd Busick: What’s up everybody? You’re listening to The Savvy CIO by Park Place Technologies. I’m your host, Bradd Busick, and this is a show where we talk about budget pressure, AI adoption, security, and the art of keeping the lights on without burning everything down, with the people who are tackling these problems every single day.

Talk to any CIO and they’ll tell you they’re all too familiar with the tension between keeping aging infrastructure running and the push to stay on trend with digital transformation. Let’s just look at the first part of that equation first. What’s the right choice to make when your OEM warranty expires? The stakes are high and the wrong choice could cost you way more than just money. Do you make room in the budget for a costly refresh? Do you bridge the gap with third party support, or is there another option? Rob Brothers has spent years thinking about this exact problem. He’s a tech industry analyst who’s tackled the infrastructure lifecycle market and IT asset management for decades. He’s here to share some hard won wisdom, not a sales pitch or talking point, on the playbook that CIOs need to have in their pocket in order to navigate what happens after the warranty ends.

Rob Brothers, welcome to The Savvy CIO.

Rob Brothers: Hi, Bradd.

Bradd Busick: Well, let’s kick things off here. Share a little bit about yourself and what you’ve seen in your career that can help some of our listeners navigate the question of what to do when the warranty actually ends.

Rob Brothers: Yeah, of course. Bradd, I’ve been in the industry for about 35, 40 years now. I was a value added reseller. I actually worked for a third party maintainer. I’m now here at DCs, so I’ve seen and done a lot. I’ve been on the front end of the sales cycle for new, I’ve been on the used side, and I’ve also been on the TPM side, so I’ve seen it and covered it pretty much from every single angle you can imagine. With that being said, third party maintainers and kind of that whole asset disposition piece has really come a long way over the course of the past 30, 40 years, and it’s been a really strategic avenue for CIOs and enterprises to take when they’re really looking to kind of garnish a little more capital, especially when we think about some of these AI initiatives that they want to take on. These are not inexpensive solution sets that they want to bring in and test, so looking for ways to stretch that dollar a little bit further in a secure, safe manner with somebody who’s extremely reliable is definitely something I see on a regular basis.

Bradd Busick: I mean, given that you’ve been tracking this lifecycle market for years, particularly as a VAR and then with many different hats, you’ve probably seen literally every make and model of the listeners that are currently tuned in right now. What’s the most surprising shift you’ve seen in how CIOs think about end of warranty decisions over the last few years?

Rob Brothers: Oh, wow. Yeah, I would say, let’s track it back over the past 10 or 15 years. We’ve definitely seen extended life, without a doubt. It was, back in the day we would see three years, was generally the lifespan, maybe four, but then after that, it began to expand. Now it’s closer to five, six, seven years, and to me, a lot of that was due to the fact that application technology had really kind of slowed down a little bit, right, and so the operating systems weren’t being updated as much, the applications weren’t being updated as much, so it wasn’t putting as much stress on the systems as when we first saw them come out. Right? Then, as time went on, customers are like, ” Whoa, wait a minute, this device is still functioning extremely well. What should I do after that year three or year four, or if the vendor goes end of life, and the asset is still functioning extremely well for me?”

That’s why you see a lot of customers turn towards a TPM. When I first started here at IDC 15 years ago or so, it was about 30 or 40% of enterprises used a TPM, and they only used them for about 30 or 40% of their environment. Right? Now, today, we see 50 to 60% of customers will use a third party maintainer, and use them for 40 to 50% of their environment.

Bradd Busick: I want to rip the bandaid off and kind of start with this idea that every CIO, including myself, dreads, which is, the OEM sends the notice that support’s ending. What triggers in my mind is, all right, I’m somewhere between pissed off / anxious, because I now have to go to my CFO and ask for another blessing of money, knowing that he or she is going to say no because IT is already, quote- unquote, too expensive. With the demand for things like AI initiatives just consuming not only the boardroom, but also the bandwidth of staff, most of the decision- makers that are listening right now are trying to balance those tight budgets with not enough talent. How much of a budget and skills gap do you see and should you expect to absorb when an OEM pushes for a full infrastructure refresh, or is it actually a feasible solution these days?

Rob Brothers: A lot of the studies that we’ve done recently show that customers are really trying to use budgetary dollars for AI initiatives, and so, you know as well as I do, that doesn’t leave a lot for everything else, and enterprises now really have to turn to somebody else to assist. That skill shortage thing is huge, because I want to take the skills that I have, right, and get them on the AI train or into these new initiatives, and we’re definitely seeing folks aging out. God, try to find somebody who knows a Z series IBM system, right, or some flavor of Unix or VMS, or you name it. There’s not a lot of these folks around either, so this is where you can turn to outside resources other than the vendor to come in and really assist, because the technical engineering capabilities that they bring to the table are excellent.

The market wouldn’t be as big as it is today for third party guys or outside resources like this, we wouldn’t be talking eight, nine, $ 10 billion market space, if they weren’t reliable, because you know that if they do one wrong thing, right, that’s their reputation. Their reputation pretty much precedes everything, in my opinion, and that’s why, if you take a look at their engineering talent and the capabilities that they have, contract management, that ease of doing business, those are all really key components that enterprises are looking for, and it’s out there. Right? They don’t have to rely upon the vendor to this caliber of resource.

Bradd Busick: Yeah. I love that you hit on some of the other reasons that a CIO would even think about a third party maintainer. I’m thoughtful now more than ever that with the budget pressures that CIOs are facing, I don’t have the luxury to go out and hire 20 people. I also don’t have the luxury to have that surgical guy that only knows the Windows 2008 R2 box, and I love that you focused on this idea of relationship and reputation.

Rob Brothers: And ease of doing business.

Bradd Busick: Ease of doing business. It all comes down to those things, and I think you’re seeing firms move away from transactional into relational, with execution.

Rob Brothers: Yeah. 100%.

Bradd Busick: Has that been true in your experience, with your customers right now?

Rob Brothers: Yeah, 100%. It’s all relationship driven instead of transactional, because they want a partner that they can work with through a lot of these issues. Right? ” Hey, how can you help me here? How can you help me there?” They’re looking for the expertise. They’re also looking for flexibility, contract flexibility, right? ” Hey, can I pay monthly? Can I pay yearly?” A lot of vendors are a little more strict in their terms and conditions and the way that they structure their contracts. That’s where I found I had the most success, was during the contract negotiations, being able to be nimble with my contracts. I can adjust that on a monthly or quarterly basis. If you remove it, you’re going to see that price differential in your next bill, which is huge, and I’m going to keep your asset inventory up to date, and I’m multi- vendor. I don’t care whose equipment you’ve got, and I’m going to be agnostic. Right?

I’m not going to try to sell you on something. I’m going to be there for what you need. Now, I’ll lend my expertise and experience and say, ” Hey, you may want to use this,” or, ” You may want to use that,” but it’s not driven by a vendor.

Bradd Busick: Well, I don’t know about you, but Rob’s got me sold on this idea that at the end of the day, when your warranty ends and you really just want a partner who’s willing to meet you where you’re at, if they also have the credibility to make it dance, that’s a pretty big competitive advantage. Because David Kramer is the president of service delivery at Park Place Technologies, his whole job is to basically think about this exact thing, so we asked him, what does the ROI case actually look like for enterprises that extend the life of their hardware versus those that undertake a hardware refresh?

David Kramer: There’s a couple variables that come into play when we’re looking at the return on investment for extended life of an IT asset. The first is a pretty obvious one, and that is, you’re delaying capital expenditures. With a lot of customers, what we see happening is the technical teams will get excited about the next generation of server or network or storage or integrated infrastructure, and they will begin a campaign, almost, internally, trying to get buy- in so that they can buy the latest and greatest technology. Usually that campaign is aided and abetted by the salespeople from the OEM itself, because the OEM is incented to sell more hardware. On the front end of this cycle, you have to take an approach where you know which hardware you want to extend the life of and you know which hardware you are going to invest in the cutting edge and have the newest and latest and greatest.

On top of the CapEx, I think there’s another side that a lot of people overlook, and that is the OpEx. Your teams have to learn the new technology. You have to pay for installation and configuration of the new technology. You need time to let the new technology bake in and test in, in your systems and in your environments, and usually there’s a complex matrix of compatibility. The database needs to run on this type of operating system, this operating system hasn’t been tested on the latest and greatest hardware, so there’s a bunch of overhead that comes with these asset migrations and these transformation projects.

Now, if you want to really complicate matters, it gets very, very interesting when you begin to think through, ” Am I talking about this investment for a new data center and new infrastructure?” Let’s call that AI infrastructure, where you’re using GPUs and you’re doing a whole new type of computing, because in those situations you have some different factors that come to bear, like scarcity in the market. In most cases, the customers that we talk to who are managing a traditional data center, obviously memory right now is hard to get ahold of. There’s a supply chain constraint on memory at this moment, but for the most part, the traditional data center hasn’t had to deal with the same problems and the scarcity that’s happened on the AI side. You’re trying to get ahold of the latest and greatest GPUs and you’re trying to build these new infrastructure pools, resource pools so that you can do the work, but it’s a race, because you’re competing with some of the biggest companies on the planet to get access to that infrastructure.

Another really important point is, the price point of that infrastructure is significantly more than the old data center. Hardware maintenance, you rarely had a $ 100,000 piece of equipment. Now, all of a sudden with these GPUs, you’ve got 100, 200, $300, 000, they call them sleds. It’s actually like, think of it like a shelf that you would slide into a dresser drawer or something. It’s a sled that has three or four GPUs that cost $ 200,000 each. An analogy I like to use is from my days as a PC gamer. Back in the old days when I had a lot of free time, I was building my own PCs, and I would regularly run out to the local hardware store and buy a new graphics card, a new GPU. I would probably do this twice a year so that I could play the latest games at the fastest speeds.

Well, the behavior that we see in the AI infrastructure space is very, very similar. As Nvidia and other providers are out there creating new chip sets and delivering new semiconductors, customers are swapping out cards and then swapping out whole sleds of GPUs inside the chassis, making the importance of maintaining the environment that much higher, but also making the complexity that much more, because they’re constantly on the IT treadmill, as I like to call it. There’s just a stair and another stair and another stair, and you’re constantly chasing the latest and greatest technology, so big changes in the environment and the industry for sure on that front.

Bradd Busick: Thanks, David. Now, back to my conversation with Rob. So much of this is not just the practicalities and logistics to calculate. There’s some emotion behind it, too. I mean, the risk of something going wrong or the risk that you could be taking on by going with the wrong provider has massive implications when you’re talking about your data center being in the wrong hands. Do you have these stories you’ve seen in your journey where a customer just made the wrong decision or rushed into a refresh they couldn’t afford, or even going the other way, of stretching their hardware too long?

Rob Brothers: Yeah, there was one they rushed into a refresh, and this one is sad but kind of funny. I was the maintenance provider for a lot of their older assets, and the VAR came in and they wanted to sell them a new storage array, and sold them this great new storage array, but the expertise and the planning of it was done really poorly, so poorly that when they went to go turn this new array on, the UPS almost exploded, because they didn’t take into consideration that the power was going to trip over the UPS capabilities of the system, right? It was really poor planning, so much so that they actually brought us in because of that multi- vendor capability, where we’re able to look at it and say, ” Okay, hey, let’s switch these things out, let’s move this around,” and we were able to get enough power to that device for the customer.

That was a forced refresh, because what they had there was actually fine, but it was going into support. It was holding some mission critical things on it, so hence they wanted to move it to vendor support, but when they found out all this happened through the vendor, they were not happy at all.

Bradd Busick: I bet. I can imagine.

Rob Brothers: I think there was two days of downtime in there.

Bradd Busick: Yeah. You just said it. That just gives me PTSD, this idea of downtime because of poor planning. I think what it highlights for the CIOs, all of which who either are driving off the road because they’re having a physical reaction to two days of downtime in their own environment, but having a partner that’s willing to be there, boots on the ground, helping dig out, not only in the people and the process, but also in the technology. Give me a little color about how product availability needs to play into planning and decision- making. It almost feels like AI is almost forcing, ” Hey, you need to have this X, Y, Z stack in order to accomplish your AI goals.” That lead time feels lengthy to me. What are your thoughts on that?

Rob Brothers: I mean, that’s super interesting, and one that I’ve been really thinking about probably just over the past three months. When we look at gen AI, and I had this conversation with a good friend of mine who’s really good in the AI space, and he said, ” Rob,” he goes, ” Look, most of the queries that we do do not need to be token- based.” Right? A lot of what we’re doing are basic searches on systems, and what we’re not planning for is, what is actual training? What actually needs tokens and what actually needs these resources? Because what we’re seeing is this little bit of a striation of everybody getting all these tokens and they’re going through millions of dollars of tokens, and when you look at what they’re utilizing them for, they can be done with simple queries on Google searches, or just better regular AI, right, not generative AI, so we are beginning to see a little bit of that.

We are definitely thinking there’s probably a little bit of an AI bubble, because if you look at the CoreWeaves and the amount of money that’s being spent in this space, we do feel that, yeah, it’s crazy. It’s not sustainable. It’s not sustainable. I mean, just the power utilization by these systems and the amount of money being invested, there’s not a return on investment yet. I mean, you’re talking about some investors that have got some pretty deep pockets.

Bradd Busick: Agreed.

Rob Brothers: With that being said, at some point in time you need an ROI. If you look at the CoreWeaves of the world, they’re not seeing a return on investment yet for all the infrastructure that they’ve put in place. We do need to move fast. Don’t get me wrong. I think generative AI is fantastic. There’s some great use cases out there. One of the recent studies we did, it was about 45 to 50% of enterprises are now beginning to see an ROI on their generative AI initiatives, which is great.

Bradd Busick: Wow.

Rob Brothers: Yeah. A year ago, if we had this conversation, it was 10 to 20%. It was a super-

Bradd Busick: Yeah, agreed.

Rob Brothers: We’re now beginning to see companies utilize this technology and be able to say, ” Hey, I’m seeing an outcome,” but with that being said, you’ve still got to keep the lights on with all the other assets that you have in your space, and those assets will be utilized for things like inferencing. There’s absolutely no reason why you need a $ 200,000 liquid cooled Neptune system from Lenovo to do inferencing. Right? These assets, I mean, they’re going to be useful, and over time, I see them doing more than just generative AI workloads. I see them doing consolidation of data centers, right? Kind of like when VMware first came out, we were able to consolidate 10 servers down to one. I think you’ll see that kind of consolidation again, but that will take some time to kind of work out.

Bradd Busick: How do you coach a CIO in order to think about the right choices that should be made for their enterprise? If you came across somebody that had aging hardware, either by choice or necessity, what would, quote- unquote, good third party support actually look like in your lens?

Rob Brothers: Well, I mean, the aging technology still has to be there, right? Because generally, when we’re talking about generative AI initiatives or new initiatives, running your existing infrastructure on the backend, so you want to make sure that you have somebody that you could turn to that can definitely run that infrastructure. From a gen AI perspective, if you think about, ” Okay, as a CIO, where should I be going?” Obviously, you’ve got to pick about three to six use cases internally that you can then go through and say, ” Hey, which ones are going to be of value?” You’ll probably get 50, right, because you’re going to turn to your internal people and you’re going to say, ” Hey.”

Bradd Busick: The floodgates open.

Rob Brothers: Floodgates open. ” Let’s just bring it on. Give me your ideas.” You’ve got to decipher down through all those. You’ll probably pick six or so, and out of the six, maybe two will pan out and create the outcomes that you need. Then, here’s the thing, testing that is interesting, right, because if we’re testing it in say, CoreWeave, it’s different than bringing it and running it internally, because there is a lot of infrastructure that’s got to be put in place in order for that to happen. Now, there’s rear door heat exchangers. There’s a lot of different directions you can go with the technology to kind of bring in high density systems. Up to, I’d say, 75, 60 to 75 kilowatts per rack, you can go with non- liquid cooled technology, but once you kind of get over that 75 kilowatt per rack threshold, direct to chip liquid cool is really the only option, either that or immersion cooling, which is a pretty interesting technology.

Yeah, I mean, strategically, you’re probably looking at bringing in somebody who understands the processes within your company, to be able to say, ” Hey, what should I be doing here? What should I consolidate down?” The biggest hurdle to all this is the data structures and the data sets. The data is generally what bogs everybody down. It’s the same thing that bogged us down in AI, is the same thing that’s bogging us down in gen AI. Good quality data and information. I’m sure everybody’s heard the crazy stories of law offices using ChatGPT to write briefs that are citing things that just don’t even exist. I mean, I think the next three years, I would say even two years, is going to be fun.

Bradd Busick: That’s exactly right. There are problems that I don’t think we even understand fully yet, and we need to start surfacing some of this so that CIOs don’t get caught going, ” Hey, I need a $ 2 million capital budget for this data center refresh,” just so you can go light up AI for this specific capability. I love your lens of, you need a partner that does more than just hardware. If our listeners have taken away anything from our time today, that right partner is out there. You’ve just got to be able to ask the right questions. Rob, it has been a pleasure chatting with you today. Thanks so much for joining us on the Savvy CIO.

Rob Brothers: Great. Thank you very much, Bradd. Appreciate the time.

Bradd Busick: Here’s the key takeaways that spring to mind from my conversation with Rob. First, finding a partner that actually can help plan, execute, and measure is the difference between a CIO having a job and a CIO not having a job. The second is showing up when they’re needed versus just selling and bouncing. Anybody can have a transaction, but I think it’s a true partner that actually sticks with you post- sale, and is with you not only in your current refresh, but the next refresh, two, three years from now. The best CIOs that I know surround themselves with partners that care just as much about their mission as they do.

That’s all for today. Thanks so much for listening. Please follow us so you don’t miss an episode. This has been The Savvy CIO brought to you by Park Place Technologies. If you want to learn more about Park Place, go to www. parkplacetechnologies. com. Now, one last word from our guest, Rob. This show is called The Savvy CIO. What is your savviest choice you’ve made in your career to date?

Rob Brothers: Oh, wow. I’ll tell you, the savviest thing was becoming an analyst. I really do enjoy having conversations with all the enterprises that I talk to, the vendors, just the entire IT industry. It’s been a lot of fun, because I get to see it now from many, many different perspectives. I think that was probably one of the smarter moves I made, as I get older.

Bradd Busick: You’re not getting older, Rob, you’re just getting wiser. I love that answer. Thanks for sharing. I appreciate it.

Rob Brothers: Thank you.

Bradd Busick: I’m your host, Bradd Busick, and as always, IT shouldn’t just be at the table. IT is the table.

Rob focuses on worldwide support and deployment services for hardware and software and provides expert insight and intelligence on how enterprises should be addressing key areas for infrastructure hardware, end user hardware, software and services. He also has expertise in the latest asset financing options which include as-a-service models such as device-as-a-service. Mr. Brothers draws on decades of industry experience to provide insights and actionable advice to assist vendors in developing, marketing, and delivering IT services. Rob’s areas of expertise include, but are not limited to, enterprise storage, server, networking, client device and high-availability and mission-critical support services, virtualization and transformational services and channel and partnering strategies.