AI Is Real. The Infrastructure Strategy Has to Be Real, Too.

By Scott Willis, President and CEO, DartPoints

At ITW 2026, you did not have to listen very long to know where the industry conversation was headed.

Power. Deployment timelines. Future capacity.

Those topics were everywhere, and for good reason. They are shaping the next phase of growth for the data center industry. But they are also creating a more complicated operating environment than the one we were working in even a few years ago.

AI has had plenty of promotion around it. That is how most major technology shifts start. There is excitement, then hype, then the harder work of figuring out what is actually real.

In this case, AI is real. We are beginning to see it materialize in practical ways. The opportunity for the industry is significant.

But opportunity is only one side of the story.

The harder question is how the industry responds. For data center providers, that means thinking beyond whether we can add capacity. It means thinking about the ecosystems we are building inside our data centers and whether those ecosystems will still be relevant as customer requirements change.

That is the work in front of us.

AI Is Changing the Conversation, But Power Is Changing the Map

DartPoints was built around a strategy focused on non-tier-one, non-primary markets. That has been part of the company’s growth thesis from the beginning.

For a long time, that may have sounded like an “edge” conversation. I have generally avoided that word because it means different things to different people. What we are talking about is more specific than that.

We are talking about where customers need infrastructure, where power is available, and where capacity can be deployed in a way that matches the timing of the demand.

That is becoming more important as AI moves from concept to deployment.

The International Energy Agency projects that global data center electricity consumption will more than double by 2030 from roughly 415 terawatt-hours (TWh) in 2024 to 945 TWh by 2030, with AI as a major driver of that growth. That does not mean every forecast will be perfect, but the direction is hard to ignore: power is becoming one of the central constraints in infrastructure planning.

Inference AI, in particular, is creating more need for geographically dispersed infrastructure. JLL’s 2026 Global Data Center Market Outlook notes that inference workloads could overtake training as the dominant AI requirement as early as 2027. 

That matters for how capacity gets planned.

Not every workload belongs in the same handful of primary markets. Some capacity needs to be closer to where customers operate, where data is created, or where latency, resiliency, and regional access matter.

Power is also forcing a different conversation.

Historically, a lot of data center growth concentrated in primary markets because that is where the infrastructure ecosystem already existed. Today, many of those markets are constrained. Power is harder to secure. Timelines are longer. The competition for capacity is intense.

That pushes more attention toward secondary markets where power may still be available and where the right operator can bring capacity to market in a way that fits the customer’s timeline.

That is one of the biggest factors shaping how we think about deploying capital at DartPoints.

The Job Has Gotten Harder

This is a great industry. It is also a harder industry than it used to be.

Looking back, the work was more straightforward in some ways. Workloads were less demanding. Growth patterns were more consistent. You could make certain assumptions about space, power, cooling, and customer requirements that are harder to make now.

That world is changing.

Today, an infrastructure provider has to think across the full environment. Cooling. Space. Operational efficiency. Utilization. Power. Connectivity. Capital deployment. Customer requirements today and several years from now.

None of those decisions can be made in isolation.

Colliers describes the current data center market as one increasingly shaped by “power scarcity, rising capital intensity, and infrastructure-scale execution.” Core markets face severe power shortages, which is forcing operators to look at secondary or tertiary geographic regions. That lines up with what many of us are seeing in the field. The demand is there, but the path to support that demand is more complex.

If you add capacity without thinking through cooling, you create a problem. If you focus on power but miss the connectivity requirements, you create a different problem. If you design only around today’s workload profile, you may limit what the facility can support later.

That is what makes this moment challenging. It is also what makes it interesting.

The providers that do well will be the ones that can make disciplined decisions, invest in the right places, and operate with enough flexibility to support where workloads are actually going.

Regional Markets Have a Larger Role to Play

The next phase of data center growth will not be limited to the largest primary markets.

Those markets will continue to matter. Hyperscale campuses will continue to matter. Large training environments will continue to matter.

But they are not the whole story.

Enterprise workloads, AI inference, hybrid infrastructure, regional resiliency, and high-density deployments are creating demand in more places. Customers are looking for practical options. They need capacity. They need power. They need credible timelines. They need an operator that understands the complexity of the environment, not just the square footage.

That is where regional and secondary markets can become more relevant.

The point is not that every workload should move out of a primary market. That would be too simplistic. The point is that customers need more choices, and power availability is becoming a much bigger part of the decision.

For DartPoints, that aligns closely with how we have thought about our business from the beginning.

Capacity Alone Is Not Enough

There is a lot of focus right now on who has capacity, who can get power, and who can move quickly.

Those things matter. They matter a lot.

But capacity by itself does not solve the problem.

The more important question is whether the capacity is part of an operating environment that can support the workload. That requires the right design, the right cooling strategy, the right connectivity, the right operational discipline, and the right capital plan.

It also requires a clear understanding of the customer.

What do they need today? What will they need in two or three years? How much flexibility will they require? What constraints are they trying to avoid? What does success look like after the deployment is complete?

These are not abstract questions anymore. They are becoming central to how infrastructure decisions get made.

How DartPoints Is Looking at the Next Phase

At DartPoints, we see a strong fit between where the market is moving and the strategy we have been building around regional infrastructure.

We are focused on markets where power, location, and customer demand create a meaningful reason to invest. We are looking at capacity through the lens of enterprise requirements, AI-driven demand, and the practical realities of deployment timelines.

That does not mean chasing every opportunity. In this environment, discipline matters.

The data center industry has a lot of tailwind right now, but tailwind does not remove the need for execution. If anything, it raises the standard. The companies that succeed will be the ones that can match demand with infrastructure that customers can actually use today and grow into over time.

That is where our attention is.

AI is changing the infrastructure conversation. Power is changing where that conversation happens. Regional data centers are going to matter more and more.

The opportunity is there.

Now the work is to build and operate in a way that can actually support it.