The Data Center Opportunity Is Shifting From New Builds To Modernization

While AI-driven hyperscale builds grab the headlines, the larger and more actionable opportunity for MSPs, VARs and systems integrators lies in the existing installed base, according to Mike Crosby, senior industry advisor at Circana.

The dominant narrative in the data center market has largely centered on AI-driven hyperscale expansion and the construction of new facilities. While those investments continue to attract headlines, the larger and more actionable opportunity for most solution providers such as MSPs, VARs and, systems integrators actually resides within the existing installed base, according to Mike Crosby, senior industry advisor at Circana.

[What follows is a transcript of the above video]

Jennifer Follett, vice president of U.S. content and executive editor, CRN: Data centers are in the news. There's so much opportunity being driven by AI and the compute power that sits behind it. I'm Jen Follett here with Mike Crosby of Circana. Mike, how are you today?

Mike Crosby, senior industry advisor, Circana: I'm doing great, great seeing ya.

Follett: Good to see you. Let's talk a little bit about this data center opportunity because it's very large, and there is a component of it that's relevant for the channel. Where do you see that biggest opportunity right now?

Crosby: It's certainly top of mind. I'm spending a lot of time with a lot of different clients really talking through the TAM [total addressable market] and the SAM [serviceable addressable market]. Really, how big is it? And then ultimately where does the channel play? In the media, you're hearing so much about how many hundreds of billions of dollars are being spent by hyperscalers or neoclouds and where all that compute is going. The question though really is where and how does the channel basically play within that environment? And unfortunately, a lot of the hyperscale spend that's being made, much of that is being fulfilled or managed on a direct basis. So what we really wanted to do is understand more what is the real opportunity, I think, for MSPs in the channel. And if you do, you kind of break it down a little bit by type. You can look at on-prem, and certainly there's a lot of channel participation with enterprise and on-prem. You also see some benefit through co-location and also on edge. That's really where the opportunity is to play, or hyperscale tends to stay outside that norm a little bit. But as you get within the segments that I was just talking about, MSPs have a pretty significant opportunity for growth, not only near term, but also over the period of these refreshes that are continuing to go on. And I know we're going to talk a little bit about it: Modernization is a key thing that I'm hearing more and more around the themes of this, and [is] certainly a great opportunity.

Follett: Let's talk about that a little more because obviously a lot of the data center headlines right now are around new data centers that are going in and do communities want them, do they not want them? How much water and power do they use? But at the same time there's tons of compute power already sitting in data centers. So where do you see the big opportunity for the channel partners in particular?

Crosby: And you nailed it right there in the question because I think what we see and hear is the the increase in regulation and the timing, permitting, all the other things that are going on relative to new builds. If you look at the existing builds, there's 5600, 5700 data centers already active, and about half of them right now are more than 10 years old. And why that's important is clearly when you build a new data center, there's new deployments of hardware and all the technology and all the services that get wrapped around that for a new data center.

If you look at half or greater than 10 years old, many of those need to be on refreshes typically between three and five years. So there's a significant opportunity where you're just going to be continuing to be on this refresh or modernization cycle for the period these data centers last. And what's interesting is most data centers, the physical building will last anywhere between 20 and 40 years. So you think about that if you're doing a significant refresh three to five, three to five, and a longer tail that has to do with ongoing maintenance and recurring revenue for the channel, it's a pretty significant opportunity.

Follett: What's the advice to channel partners on how they should start to go after this opportunity and what the different pieces are that they really should be focusing on?

Crosby: Well, and the best part about I think where we are is not only the quantity that we were talking about, but also as you get to midsized business or smaller businesses, they typically don't have the existing infrastructure to manage the daily care and feeding, nor the planning, nor the refresh. So this is where they're going to augment whatever they have with channel resources. And that's where the channel partner can really play a pretty integral role in this.

And one of the things that we started to see more than ever is through assessments. Channel partners or MSPs can come in and they can do an evaluation and understand where are you? Where are you in the life cycle? What's the requirement? And build it an onboarding plan for the refresh of the technology, plus that longer tail I was talking about around these recurring services. So there's there's a lot of really consultative elements to the MSP engagement with these. Much of it is an explaining, again, telling them where they are and where they are from a productivity standpoint, from a capability standpoint, but even more than that, laying out a strategy and a plan on how do you continue to modernize over the periods of time.

For a 20- or a 40-year-old building, in many cases outside the core—being cooling or power—you're going to see that every three to five years, there's opportunities for that to be an ongoing refresh kind of a plan. And again, it builds on a relationship. And I think that's where the channel has always been so great at. And that's the capability that I think really gives them, I think, a real leg up.

Enterprise, the larger businesses, in many cases, while there's still opportunities for MSPs, in many cases they have their own existing infrastructure. So it likely won't be the level of augmentation nor the ability for the MSP to really help consult. And I think that's where a critical part's going to be. And that bridges the gap between not only the hardware, but the software, the services, the security, all the different elements that we're talking about relative to ongoing updating of these facilities.

Follett: What does it mean now, today, to say that your infrastructure is AI-ready?

Crosby: Yeah, it's a great question because a lot of people think AI-ready is just, well, you're going to upgrade the server or you're going upgrade some of the technology. And it really is looking at all those core elements along the stack. You really have to make sure, because AI is so much more intensive relative to that, it's looking at each aspect from compute, storage, security, networking, power. All of those elements need to be not only identified and scoped, but ultimately making sure that all those are firing at the level they need to be to be able to support the added burden that AI is going to bring to the facility.

So there's a lot of that. It's not just a server upgrade, there's significantly more than that. And that also again bodes well back to the overall planning. Not only where we are today, but ultimately where we need to take it. We go to the end result and work backwards. And that's where the MSP really has a great opportunity to really model this and have it set up the way it needs to be set up, not only for today but the future.

Follett: What do you think are the elements of the infrastructure stack that are most out-of-date that the solution provider should really be focusing on?

Crosby: You know, the tough thing is right now they all work kind in in concert with each other. I would probably lead with security first. But then beyond security, I think you've got to look at all those as an integrated solution. I think it gets very difficult where you're upgrading one element or looking at one key element if it's just compute or if it's just whatever else. Much of that needs to be managed, but that umbrella of security tends to be the one that is the cornerstone. And then beyond that, it's the rest of the ability to process and power and cool.

Follett: And then what do you feel like are the biggest roadblocks or hurdles that they need to get over as they're trying to have these conversations?

Crosby: You have to look at it from a couple of perspectives. One is in many cases, the hardware is the opportunity initially, but I think the real margin proposition for the MSP is really on the integration services and the ongoing maintenance and monitoring and the other planning that goes with that. And so I think what we have to get them thinking is less horizontally and think more vertically. Some of the real benefits I hear is if MSPs are digging deeper specifically into a vertical like security and really going to at the expertise and at the level that's needed to really be able to consult within those areas. So I think areas of specialization are really going to be critical for the MSP. And I think they have to understand where they have the resources, where they have the expertise, and how do I leverage that accordingly.

Same thing with looking at segmentation. If they have a background in health care, in banking and finance, in other areas, utilities. These are things where you can align those to more of that vertical application and that environment. and that tends to elevate margins and also tends to protect them a little bit relative to competition because it's something that they can take advantage of and exploit.

So I think part of that is think less about just horizontal capability and begin to understand what are those verticals to attack, but also what are those requirements to support that vertical. And then you build backwards within the infrastructure. You have to make sure you've got the expertise that's warranted for that type of a solution. But specialization is great because it protects them competitively, but it also brings, typically, incremental margins as well.

Follett: Seems like power management and cooling are having a moment here because of the tremendous compute power that's being drawn to do all these amazing AI things, but can you talk a little bit about how you see that power focused opportunity playing out with the channel partners?

Crosby: I think you're exactly right because these are the two big pressure points that we're getting from a lot of the communities now on having a data center in my environment. How are you going to power it? Is it going to add expense and cost to me because you're tipping into the grid? Are you generating your own power? Are you also looking at cooling where maybe it's not water cooled, but it's other different technologies that are emerging? These are really critical areas that I think are real opportunities for MSPs to delve deeper into that expertise because in many cases, the legacy install that was done with the originating build of the data center, we've already seen technology evolve. And that bridge between old and new is a really critical area because they're getting more efficient as we see it. They're using less power, they're taking less to cool, they're using new technologies and new means to do that. This is where I think the MSP can get on the forefront. And again, that's real leverage, especially looking at a legacy install of a legacy data center that maybe has a 10-year period on them, so I think you're going to see those two areas specifically really heavily exploited by a lot of MSPs that that really see and can embrace the opportunity.

Follett: What about for a solution provider that really wants to be making sure they're having a strategic data center conversation, what are the software pieces that they need to be including in that talk?

Crosby: Clearly, and as I'd said earlier, security is the first part, is really making sure that they've got a core knowledge, again, enabling them to really understand not only the needs and requirements of the data center today, but more importantly, if it's within some of these different verticals, like we touched on—health care, banking and finance—where there are unique requirements, because a lot of these are very highly regulated industries as well. So knowing the ins and outs, now again, not just horizontally, but I would say security is definitely the lead pin, but also understanding the dynamics for each of the verticals they're playing in, that really sets them apart from many others that are just talking more either hardware on a refresh or basic hardware or software install. It really takes that to the next level.

Follett: Flipping back to the hardware piece though, so much of this is dependent on supply and what's actually available. So what's your take right now on how the supply situation is shaking out and how many of these systems are even readily available right now?

Crosby: It's a difficult time because, as we talked about before a couple of different times, specifically around DRAM and NAND, memory and storage continue to be challenged. As AI has done so well, the challenge with the suppliers is being able to keep up with that. So a lot of things that we're seeing now is working very closely with additional partners, if it's distribution or other partners, making sure that you can have reserves available or you can do some different things where in many cases you're getting end users or pre-buying and then scheduling out the allocation of the technology. So a lot of it is working much closer with their fulfillment partner and knowing and understanding what likely is going to be the availability and then getting commits from the end user sooner. As I'd said, there's the ability to reserve, and we've seen a lot of that take place.

But to your point, there's still going to be some challenges. There's longer lead times. We are still not anticipating through the end of likely ’26, certainly through ’27, for either memory or storage, we're likely to see storage improve just a little bit. And they'll start with availability, and likely pricing will start to normalize as well. From a RAM standpoint or DRAM side, we we're seeing that likely going to continue for a little bit of time. So you're absolutely correct. I mean, you can build or you can plan for what's needed and required, but you're going need to make sure you're building in adequate levels of lead time because there are going to be some tightness around supply. And that just going to be leaning to the planning and maybe making sure that's done at the appropriate time.

Follett: You did mention neoclouds at the top, and some of those providers are expressing interest in working with channel partners, so I'm just curious how you're seeing that shake out.

Crosby: I mean neoclouds, as we've seen and talked [about], it's really GPU as a service. It's really the ability to, on demand, really augment what's needed. And I think there's a real need there. I think the question is, I think it's still going to be access to product that they need and access to GPU. So I think you're definitely seeing not only neo clouds look at what my access to product is directly, but also how do I take advantage of channel partners and relationships and additional inventory that may give me the ability to bridge that a little bit. What's interesting too is even with most neoclouds, they're not really necessarily driving their own facility. They're still in colo facilities and other ways that are leveraging off existing infrastructure that's already built. But I'm providing the GPU capacity that ultimately makes that happen.

So they're getting creative on their side as well and leveraging hardware and leveraging access to hardware and access to facilities that ultimately enable them to do what they do. This is such a dynamic time where the demand is by far outstripping supply, and you have to get very creative in a lot of different ways that you're able to access the critical needs to support these. And I think that's going to continue certainly for the next, easily the next couple of years. And I think flexibility on that is just going to be the key.

Follett: Yeah. And we've had past conversations talking about the edge opportunity. So maybe just touch a little bit on, you know, the difference now between what what's going into the data center, what's going in the edge, and how that edge opportunity is impacting the data center spend.

Crosby: Well, and it certainly is there because of the low- latency opportunity that edge brings. And I think you're seeing as demand for those types of resources and speed and capacity and bandwidth are picking up a little bit, you're seeing that continued investment, not to mention closer to where the data is actually occurring. And you're seeing it as we see smart cities, as we're seeing autonomous vehicles continue. All of the same momentum is continuing to build, that demand is accelerating even more, and the beauty of edge is typically it's a little bit faster to deploy. In a lot of cases, that's not necessarily refreshing an existing edge location, but it's seeing new deployment that's going on because you're using less requirements from a property standpoint. There's less requirements and regulation on those. And I think that's where we're kind of finding a little bit more of a flexibility tied to these opportunities and building these new facilities. So certainly that's the case. The hardware requirements are still there, the other requirements to physically be able to deploy, but I think the regulatory part of edge makes it a little bit more available and certainly a little bit easier, but also driven significantly by demand.

We are seeing clearly more and more and more need and requirement for closer to where the data is is occurring and being captured and ultimately being able to manage these things dynamically, like I said, in those environments: smart home, certainly smart city, certainly autonomous vehicles, et cetera. Those are all going to be not just more but significantly more required.

Follett: So for a solution provider that's playing this role of strategic advisor, there's no shortage of things they could recommend that get upgraded in the data center, but that doesn't mean the money is there to do all these recommendations. So how should they be talking to customers about balancing investments in their AI initiatives, new technology to support those AI initiatives, and then whatever their traditional needs are.

Crosby: Yeah, and like you said, I think that's where it's the critical part of what the MSP can do is they've got to have the expertise and the knowledge to be able to do these assessments and planning, because some of this is not a one and done. This is going to be a relationship as they build, especially for multiple facilities, that they're going to want to be in part of that refresh cadence over that next three to five years. So it's going to be critical to make sure that they can do a a critical plan, not only a current assessment, where are we, but more importantly, where we are the next cycle, the next cycle and the next cycle after that. And I think that's going to be the linchpin to this is it can't be transactional, it needs to be much more consultative, and it with that is going to need to not only know and understand the technology of today, but also through these multiple cycles and refreshes, how does this evolve over time? Because there's also going to be efficiencies that come with this, that it's not just going to be the same thing over and over again every three-to-five-year cycle. You're going to see new technologies that they also need to be on the forefront of.

Keep in mind they're acting in that as that consultant, in that consultative manner, that's educating these these clients specifically on not only today, but like I said, tomorrow. So I think the critical part of that, being able to effectively plan, plan in stages, being able to build and leverage off that that plan over multiple refresh cycles. This whole modernization kind of thinking is going to be I think a huge, as we talked about, a huge opportunity because these are functioning facilities that are going to live for 20, 30, 40 years. But internally, right, there's going to have to be able to be dynamic in the way that these plans can evolve as the technology evolves. And I think that's where it's going to be the critical part of what they bring to the table is really going to be that knowledge and expertise and the depth of that expertise. It has the ability really to educate the client again and be that partner, be that partner for a longer period of time.

That's the beauty of this, also, I think, Jen, is that this really builds on the quality of the relationship and the longevity of the relationship because there's going to be a lot of trust involved after that initial assessment and implementation plan. Then it's going to be the ongoing maintenance that really carries forward. And I think that's going to be, again, I feel very, very bullish on that long-term opportunity. They're going to have to continue to invest in their resources, in their knowledge base, and really making sure they're staying on the forefront because this technology is changing so quickly. It's going to be an investment ongoing, but it'll pay off. I certainly believe that.

Follett: So what's the piece of advice or the lasting thought you really want to leave our viewers with as we talk to them about making sure that they are primed to capitalize on this AI moment, particularly as it relates to the data center?

Crosby: Yeah, I think building on everything that we've already talked about, but I think the biggest one, focus on the outcome. If you focus on outcomes for the client, that clearly drives everything else. And I think that's also going to drive your own internal assessment of where's our capability? Where do we have resources that are available? And where are we meeting or exceeding what the requirements are of the client? And I think that's going to be constantly doing the internal auditing and investing and developing as a whole. But most important, as I'd said, staying focused on the outcome.

So many times I have conversations with so many clients and many OEMs too that I think they get hung up on the “what.” What are we doing? How does it work? The technical requirements that are there, etc. But most importantly, if you can't lose sight of the outcome for the client, what is really the critical need that they have that we need to continue to be solving? And then we work backwards from there and reinforce our resources early around the outcome.

I think that's a really just— it sounds simple, but so many times I think people get distracted from staying focused, again, on what is the real critical outcome for the client. And I think that's it, that really makes it a winning solution or not.

Follett: Some good ideas to build on, Mike. Thank you so much for joining me today. Really appreciate it. Had a great convo. Thank you.

Crosby: Great, appreciate it. Thanks for the time. Great seeing you.