AI is changing far more than demand for data centers. It is exposing the limits of the way our industry has designed, procured, built, and delivered infrastructure over the past two decades.

For years, scalability was relatively straightforward: build more capacity, secure more power, and expand into new markets. Success was largely measured by how quickly operators could speculate and add megawatts.

AI has fundamentally changed that equation

The challenge today is no longer simply to build larger campuses. It is to deliver hundreds of megawatts of AI-ready infrastructure at unprecedented speed and with consistent quality across multiple geographies, the need for high degree of predictability in capacity delivery dates, while integrating new cooling technologies, managing increasingly restricted supply chains, and meeting rising sustainability expectations.

These are no longer engineering challenges alone. They are supply chain integration challenges.

I believe this requires us to rethink one of the industry’s longest-standing paradoxes: how to preserve consistency while adapting to change. Delivering data centers at scale demands repeatable architectures, yet every geography, regulation, and technology generation introduces new requirements.

That model has served the industry well for many years. But as AI accelerates demand, the question is no longer whether we can design an exceptional facility. It is whether we can deliver exceptional facilities repeatedly, predictably, quickly, and at industrial scale.

In the years ahead, simplicity will no longer be just a design philosophy. It will become the foundation of scalability.

Technology is moving faster than infrastructure

The challenge is no longer keeping pace with infrastructure demand. It is keeping pace with technology itself.

AI is compressing technology cycles to such an extent that traditional real estate development models are struggling to keep up. Data centers must now be designed for the technology that will exist at the point of RFS, not for the technology available when the first drawings are produced.

This tension between consistency and adaptability is becoming one of the industry's defining challenges. Future success will depend on the ability to deliver standardized, repeatable infrastructure at scale while remaining flexible enough to accommodate rapidly evolving chip, power, and cooling technologies.

From construction projects to industrial products

For decades, data centers have been delivered primarily as construction projects. Each facility has been designed as a unique response to a specific customer, location or technical requirement.

That approach worked when growth was incremental.

But AI is changing both the speed and scale of deployment. The future of data center delivery will increasingly resemble advanced manufacturing: designing a platform once, improving it continuously, and deploying it repeatedly.

This requires a fundamental shift in mindset: moving from projects to products.

A product approach does not mean creating identical data centers. Markets remain different, with varying regulations, climates, grid conditions, and customer expectations.

Standardization should not eliminate flexibility; it should focus it where it creates real value. The objective is not uniformity. It is identifying what genuinely needs to change, and what does not.

The “Adjust to Suit” principle provides a practical framework. The majority of a data center’s engineering, from electrical architecture and cooling systems to modular infrastructure and delivery processes, can be designed as a repeatable platform. The remaining elements can then be adapted to local requirements without compromising efficiency or quality.

That is the essence of moving from projects to products.

The “Adjust to Suit” advantage: making repeatability a strength

Repeatability creates value at every stage of the lifecycle.

Every element that can be repeated becomes easier to manufacture, easier to procure, easier to install, and easier to operate. Standardized components also benefit from accumulated operational experience, allowing lessons learned on one campus to improve performance across every subsequent deployment.

The result is not simply faster delivery. It is higher quality, greater predictability, and lower operational risk.

Product thinking also changes the way we engage with customers.

For many years, customer centricity has often been equated with offering unlimited optionality. Yet not every option creates value. Some requests genuinely improve performance or address a specific operational need. Others simply add complexity without delivering meaningful benefits.

As our industry scales to support AI, learning to distinguish between the two becomes essential.

Customer centricity should not be measured by the number of bespoke features we can offer. It should be measured by our ability to deliver infrastructure that performs exactly as customers expect, on time, at scale and with consistent quality.

Standardization is therefore not the opposite of customer focus. Quite the opposite: it is what allows customer value to be delivered consistently, repeatedly, and at scale.

Simplicity as a sustainability strategy

The benefits of this transformation extend beyond delivery and operations.

Sustainability is often associated with adding new technologies or improving efficiency metrics. Those elements remain essential, but the next generation of sustainable data centers will also come from eliminating unnecessary complexity.

Every bespoke design creates additional engineering effort, supply chain complexity, material variation, and operational challenges. A more industrial approach helps reduce these inefficiencies from the outset.

Simplicity is therefore not about doing less. It is about designing smarter.

The data center industry has always been driven by innovation. The next stage of innovation will not only be measured by the technologies we deploy, but by our ability to create infrastructure that can be replicated at scale.

The companies that lead the AI era will not necessarily be those capable of building one exceptional data center.

They will be those capable of building exceptional data centers repeatedly and predictably.

Because in the age of AI, simplicity is no longer a compromise. It is the foundation of scalability.