The Idea That Arrived Too Early
A young Data Architect sits across from members of a project team led by the government to build out more infrastructure for a product that is about to grow exponentially: data.
The answer was simple: more data centers. But this is not your typical data architect.
Before Kathleen Westover’s career centered on making data actually usable, she came up through two distinct worlds: engineering and construction management.
“If you think about it, the purpose of an engineer is to find and solve problems, right? Give me a problem, I want to solve it. I found that data was a great tool for solving common problems.”
And the need for data centers for storage, accessibility, and sharing was only growing for the US government. When it came to efficiency, repeatability, and keeping costs down, prefabrication - the process of building in a factory instead of on site - was the clear winner.
But the year was 1994, and prefab hadn’t found its ideal client yet or a use case to push back against decades old public opinion.
“They said what they wanted was too unique,” Kathleen explains. “I said, no, it's not unique. If you look at our requirements, 85% of them are the same between the data centers...It was actually dismissed. I didn't get a win.”
Prefab is not a new method. The conditions were just never quite right for it to take off.
However, times have changed. Today, our hunger for data grows exponentially, and private companies are pouring billions in. The result? An ideal client and environment for a once dismissed method of building.
A Good Idea With Nowhere to Go
If you need housing up quickly, prefab is the answer. It's the reason it took off in the 1940s.
With soldiers returning from after World War II housing needs boomed. So, William Levitt applied the manufacturing methods of standardization and sequencing to homebuilding.
“There are huge advantages of building [homes] in factories," explains Kathleen. "You can control most everything in a factory environment and if you build it modularly enough, you can give people enough flexibility and variability that they don't feel like they're getting a cookie cutter one.”
For a while, it worked. But as prosperity returned and aspirations grew, the appetite for standardization faded.
That fear of ‘cookie cutter’ is one of the reasons prefab was dismissed quickly. Prefab became associated with cheapness and lower value. Zoning codes compounded the problem, making it impossible for factories to build to a single standard.
Plus, prefab completely turns construction’s fragmented delivery model on its head.
“Construction processes were traditionally waterfall, where you do each of the project phases in order. You get your requirements, then you design it, then you start building it. But you don't go back and revisit each one of those.”
But the resistance to prefab doesn’t end there, either.
“It takes an investment in infrastructure to build system-built homes,” says Kathleen. “You have to have a factory. A lot of construction companies don't have the resources to make that investment.”
And so the pattern remained: right idea, wrong conditions.
The Market That Couldn't Say No
Fast forward to today, and there are 3,069 data centers with 1,489 being built. The pace accelerated in 2022, driven by generative AI and the launch of ChatGPT.
Construction is slowing in all sectors besides data centers. Now, prefab has a market that can’t afford to say no.
Google, Meta, and Microsoft ARE the market.
But they didn’t champion prefab out of principle. They utilized it out of necessity.
When it comes to speed and demand, prefab meets the requirements. Especially when, as Kathleen noted three decades ago, 85% of those requirements remain exactly the same.
“If you owned a construction company and you were fortunate enough to get a bid to build a data center and you tracked all those costs and all that time and all the materials and all the subs — you have a very strong understanding of that particular data center,” says Kathleen.
“Then your power is, you go to the next client and say, ‘Look, I've done this before. I know what this cost, I know what the requirements are, I know that this building will meet those requirements.’”
“You can go to the next client and say, ‘Look, I've done this before. I know what this costs, I know what the requirements are, and I know that this building will meet those requirements.’”
They are the same benefits Kathleen identified in the 1990’s. Her idea from the 1990s is now the only viable method for building the infrastructure the AI economy runs on.
It’s not that someone finally listened. It’s just that the conditions finally showed up.
So What Do We Build Next?
If the data center boom was proof of concept for prefab, will the method be more quickly adopted by the industries where it matters most?
Kathleen brings up two major use cases that connect to what might be the biggest thing she believes in: get people in homes.
“Especially in areas that have had disasters, like Maui, where they had that fire. How are you going to house all those people quickly? System built.”
Disaster relief is the original reason why she studied modular building. It's a thread that runs all the way back to her graduate research. As someone who chose construction specifically to make a difference, her interest in solving real problems like this remains today.
But there is also the issue of price.
“Today, construction costs have gone from around $125 a square foot to about $300 a square foot,” says Kathleen. “Most people cannot afford that. So, I think that systems built may be coming back.”
With prefab and building advances, Kathleen believes that the preconceived notions about homes built in factories will ultimately fade.
When her community in the 1980s decided they wanted neighborhood schools, building six of them was a tall order. They built every school on a single set of plans. Because of this, the cost was known before construction even started.
The only thing they did was change the facade.
“Having repeatable processes and reusing plans that have already been developed and are well known and well understood makes a lot of sense. It saves taxpayer money,” says Kathleen.
With proof of concept in hand and prefab technology advancing, the question becomes: who will prefab’s next ideal client be?
