Manufacturing and construction teams have historically focused on supply chain performance to cut costs. In today’s data center buildout, the focus is completely different, says David Shillingford, a supply chain advisor and technology investor who’s been working with hyperscalers and suppliers who support them.

“It's all about getting stuff built quickly, and it's a different way of applying supply chain discipline and skills,” Shillingford says.

In this time-crunched environment, supply chain effectiveness delivers a huge competitive advantage, whether it’s for a hyperscaler building a new data center, an enterprise data center expanding its facility, or a data center supplier struggling to scale production fast enough to meet demand.

“The challenge of building data centers on time, at the scale that is being anticipated, that's one of the biggest supply chain challenges the world has ever seen,” says Shillingford. “The reality is that some data centers will get built and some won't. And the question then is, what's the difference between them?”

Shillingford shared his insights on two broad supply chain failures that will help determine who gets a data center built and who doesn’t.

‘Death by a Thousand Cuts’

The first source of failure comes from what Shillingford calls “death by a thousand cuts,” an inability to monitor and control the inbound logistics of all the material inputs that have to arrive, at the right time, on a data center site.

“This means that we've got this cable, this capacitor, this equipment we've ordered, and it's on its way, but we don't know where it is,” he says. “We don't know if it showed up or not. We don't know where it is on the worksite. If that’s happening once or twice or 100 times a day, that's what I think of as a thousand cuts.”

The challenge and scope are daunting, with thousands of pieces of equipment, prefabricated parts, and materials coming into a project site that need tracking. But the supply chain skills and systems required to deliver this oversight and control are well known. Shillingford is sometimes called in to help a fast-growing supplier to the data center industry improve on supply chain fundamentals, often at the encouragement of a hyperscaler that would like to buy more from a supplier if they could deliver more reliably.

“It's relatively simple supply chain execution, but you're either doing it or you're not doing it,” he says.

The Mission-Critical Component Risk

The second source of failure is harder to manage and predict: the supply chain for the big, modular, mission-critical equipment, with very long lead times, that data centers need. Think server racks, cooling equipment, or power generation and storage equipment. “If that gets disrupted, the closer it is to the final delivery, the worse the impact is going to be,” Shillingford says.

Such critical component gaps happen today for a whole range of reasons: a supplier lacks capacity, scarce equipment gets allocated to a bigger, more influential buyer, or equipment gets damaged or stolen en route. Sometimes the only way to overcome these challenges is paying a higher premium to move up in the queue. But even that wouldn’t help if there’s a global supply crunch.

“The challenge comes if there's a sudden capacity shock,” says Shillingford. “So, what's the data center version of the Strait of Hormuz?”

The answer to that question is probably something affecting chip production and shipment from Taiwan, given its dominant role in manufacturing. Such a disruption could come from any of a range of calamities, from conflict and trade restrictions or natural disasters.

“If someone said to just choose one thing that should keep data center people up at night, out of 100 things, that's probably top of the list,” he says.