The Real Infrastructure Crisis Isn’t Technical. It’s Strategic.
How economics, AI, and risk are reshaping expectations for CIOs.
Written by Alex Chang | 5 min • July 30, 2025
The Real Infrastructure Crisis Isn’t Technical. It’s Strategic.
How economics, AI, and risk are reshaping expectations for CIOs.
Written by Alex Chang | 5 min • July 30, 2025
Infrastructure is no longer just about what sits in the data center—it’s about how IT delivers it. How can IT add value, manage risk, and enable innovation?
The Array sat down with Jeb Horton, Senior Vice President, Global Services at Hitachi Vantara, to get the answers. Horton has a front-row view into how IT leaders are rethinking infrastructure. In this conversation, he explains why infrastructure strategy is being reshaped by economics, AI, and risk; why cloud isn’t the default answer anymore; and why the real job of today’s CIO is to deliver consistent outcomes, not just technology.
The Array: IT Infrastructure has re-emerged as a strategic priority. What’s changed in the past few years to elevate its importance again?
Jeb Horton: There are two big inflection points: cloud and AI.
Public cloud changed the way we consume infrastructure. It taught us to expect elasticity, scalability, and consumption-based pricing. That’s been happening for 20 years. What’s new is that people have started to realize they probably over-pivoted to the cloud. So now you see this pullback. Companies want to bring some of that infrastructure back on-prem, where they can manage cost, security and performance differently, in the way they had with public cloud.
Then you add AI, which brings a whole new level of complexity. AI needs high performance, low latency, massive data management and movement, and absolute reliability. You can't run that on shaky infrastructure. So, infrastructure suddenly becomes very strategic again, especially as people figure out how to support AI in a way that aligns with business outcomes.
TA: How are economic pressures—like higher interest rates and tighter budgets—reshaping how CIOs approach infrastructure today?
JH: These pressures aren’t new. I remember doing an ERP project 20 years ago, and the infrastructure cost ballooned by $25 million overnight because we hadn’t scoped it properly.
But what has changed is the payment model. People are thinking in terms of, “How much do I want to commit to?” and “How much flexibility do I need for the ups and downs of my business?” Again, that shift really started with public cloud, and now it’s extending into on-prem. The classic problem we all face is trying to get the best deal on the committed portion, and then figuring out what premium you're willing to pay for flexibility.
I live this, too. I'm not just a provider—I’m a buyer. I’m also trying to manage my own infrastructure spend. So I have to ask the same questions: How much do I buy upfront? How much can I stretch? What goes to cloud providers? What stays on-prem? Then I tie all of that back to my revenue model. That’s the real challenge.
TA: Tariffs and sourcing disruptions have also made this more complicated. What’s the ripple effect on infrastructure decision-making?
JH: It’s a real issue. New tariffs immediately affect how I think about sourcing, because I have data centers in the U.S., outside the U.S., and customers all over the world.
But smart customers are doing something clever. They’re locking in pricing and shifting the burden to the provider. One of my customers recently said, “Look, I want a fixed rate. Don’t pass tariff impacts on to me.” That puts the pressure on me to figure out how to manage around it, through sourcing, pricing, or procurement strategies.
That’s the essence of a good infrastructure-as-a-service model. The customer gets cost predictability, and the provider takes on the risk of figuring it out behind the scenes. Everyone's trying to work around these disruptions, and that's where the responsibility should land—on the provider.
TA: The term “Infrastructure-as-a-Service” (IaaS) gets thrown around a lot. In practice, what does it really mean?
JH: First of all, it’s not just a cloud experience. It’s a way of delivering infrastructure that matches how businesses want to consume it—on demand, priced per unit, and tied to outcomes. Think of it as cost-per-gig, cost-per-node, or cost-per-service, depending on what you're doing.
There are different flavors. On-prem IaaS lives in your data center, offering more control, compliance, and customization, which is especially valuable in regulated industries. Public cloud IaaS gives you speed and scale, but often with less control and the risk of vendor lock-in. Most companies are going hybrid to get the best of both worlds: agility with governance and cost control.
"Beyond the economics, infrastructure-as-a-service shifts accountability. "
We always talk about P times Q: price times quantity. You want to be able to manage a unit cost and scale it up or down. But beyond the economics, infrastructure-as-a-service shifts accountability. You’re not just buying hardware—you’re expecting someone to deliver performance and uptime as a service. You want to set expectations and not have to worry about the rest. And if they deliver, you renew.
TA: As you mentioned, AI is a major inflection point. What kinds of infrastructure and data decisions are becoming critical as AI moves from experimentation to implementation?
JH: AI is a massive data and latency problem. You need to move a lot of data, at the right time and very quickly, to make real-time decisions. That means your infrastructure has to be fast, solid, and scalable. But it also has to be reliable in a way that’s different from BI or analytics. If your BI tool goes down, maybe the report’s late. But if your AI model fails, your customer experience—or your entire business process—can stop. So resilience is everything.
And then you’ve got to figure out how you’re going to deliver that. Is that on-prem, in the public cloud, or hybrid? I see a lot of customers leaning toward on-prem because of performance, security, and cost reasons.
TA: Can you go more into why on-prem infrastructure is gaining relevance for AI workloads?
JH: Network latency is a huge factor. AI models often need low-latency access to massive data sets. If you can reduce how far that data has to travel, you minimize risk. That’s what on-prem gives you.
There’s also reliability. With on-prem infrastructure, you can build something rock solid that you control. Public cloud has scale, sure, but even AWS is now offering on-prem solutions. Everyone’s realizing it’s not about either/or anymore. It’s about the right mix.
And then there’s cost. Moving data in and out of public cloud environments isn’t cheap. AI workloads are heavy. The economics often point back to on-prem, especially if you're trying to optimize for both cost and performance.
TA: How does data sovereignty and security play a role in cloud vs on-prem infrastructure decisions?
JH: They’re key factors, especially for regulated industries. Data sovereignty ensures information stays within legal jurisdictions, which can be tricky in public cloud environments where data may cross borders. Security’s another driver. Cloud providers offer strong protections, but some organizations still prefer the control and isolation of on-prem for sensitive workloads.
Again, that’s why companies are going hybrid. It’s less about location and more about aligning infrastructure with trust and regulatory needs. As infrastructure shifts to a service-based model, things like trust, SLAs, and compliance guarantees start to matter more than traditional product features.
TA: What’s the biggest blind spot you still see in how CIOs or CFOs approach infrastructure—and what’s the consequence?
JH: Consistency is underrated.
From a CEO’s perspective, you want predictable costs and no surprises. You want to know your infrastructure isn’t going to fail during a ransomware attack or an outage. And you don’t want your CIO or CTO coming back every year asking for more budget without showing better results.
Too often, leaders treat IT infrastructure as a static asset. It’s not. It’s a living, evolving service. And if you don’t manage it that way, the business ends up exposed. But to get there, you need infrastructure that's not just reliable—it’s managed like a service. That means thinking beyond the tech. You’re managing uptime, resilience, and cost consistency. That’s what the CEO actually cares about. Everything else is noise.
TA: What mindset shift will define the next wave of infrastructure leadership?
JH: The next generation of infrastructure leaders won’t think in terms of what hardware they own. They’ll think in terms of what outcomes they’re delivering to the business—at what cost, what risk, and what level of resilience.
That’s where models like IaaS are powerful. They free up IT to focus on value, not just operations. And that’s what the C-suite is really asking for: Don’t tell me what you bought. Tell me what you delivered.