Not Every Growing Workload Needs a Hyperscale Answer

By Scott Willis, CEO, DartPoints.

The data center market is moving fast right now. Capital is flowing. Power is tightening. AI is reshaping demand.

But the most important shift isn’t about size or hype. It’s about how different customers actually grow — and how infrastructure needs to support that growth without forcing a one-size-fits-all model.

That’s where the industry is changing.

Growth No Longer Follows a Single Pattern

Historically, workloads followed fairly predictable paths. Smaller deployments stayed small. Larger deployments were planned years in advance. Capacity planning was linear.

That’s no longer the case.

Today, we see a wide range of customers — from local and regional businesses to national enterprises, AI platforms, and service providers — dealing with the same core challenge: their infrastructure needs are expanding faster than traditional timelines allow.

• What used to be a few cabinets becomes a megawatt.
• What used to be 7–10 kW per rack becomes 20, 30, or 40 kW.
• What used to be a “later” decision becomes an immediate constraint.

The issue isn’t customer size. It’s rate of change.

From Learning AI to Inference AI

AI is a major driver of this shift, but not just at the hyperscale level.

Much of the industry conversation centers on large training workloads — massive learning environments built by hyperscalers. Those are real, but they represent only part of the market.

What we see growing rapidly is inference AI: the application of AI inside real operating environments. These workloads show up across many customer types — manufacturing, healthcare, logistics, customer experience, analytics, and more.

Inference workloads tend to be:

• Smaller than hyperscale training clusters
• More distributed
• Closer to users and data sources
• Highly sensitive to latency and availability

That makes location, density, and time to market far more important than raw scale.

How We Think About the Market

As DartPoints enters its next growth phase, we’re intentionally designing our platform around workload behavior, not customer labels.

We support many types of customers, including:

• Customers starting with a few cabinets and planning to grow
• Regional and national organizations modernizing critical infrastructure
• High-performance and AI-driven environments requiring higher densities
• Channel-led and service-provider deployments that need clear capacity alignment

Across all of these, the common thread is flexibility with a defined growth path.

That’s why our investments focus on:

• Expanding existing sites where power is already available
• Adding capacity through brownfield and targeted greenfield development
• Supporting densities from traditional air up through advanced liquid cooling
• Prioritizing speed to market over long-dated, speculative builds

Ownership, Capital, and Optionality

Another important shift for us has been increasing asset ownership. Owning infrastructure gives us the ability to invest deliberately, densify responsibly, and adapt sites to evolving demand — whether that demand comes from a single large deployment or many smaller ones growing over time.

It also creates optionality. In some cases, that means operating independently. In others, it means partnering at the site level with investors who have land or power but need an experienced operator to bring the asset fully online.

As density and capital requirements rise, discipline matters. Deploying capital efficiently — and in alignment with real demand — is essential.

A Broader View of Data Center Growth

Hyperscale remains an important part of the ecosystem. But it’s not the only place value is being created.

Across the market, customers are looking for infrastructure that meets them where they are today — and can scale with them tomorrow — without forcing premature commitments or unnecessary complexity.

What this looks like in practice

In practical terms, this means expanding existing sites, bringing new megawatts online faster, and supporting higher-density workloads across multiple markets. 

It’s a strategy focused on execution, not speculation — aligning capital, power, and infrastructure with how our customers actually grow.