The Real Bottleneck Isn't Materials...It's Your People
The data center industry is experiencing something that feels less like evolution and more like rupture. Not the dramatic kind that makes headlines, but the quieter, more consequential shift in what clients actually value when billions of dollars and compressed timelines collide.
Tim Straub’s journey from superintendent at Whiting Turner to managing construction quality and commissioning for Google across the Americas positioned him to witness this transformation firsthand. What he describes isn’t just faster schedules or bigger budgets. It’s a fundamental recalibration of how the industry defines success, manages risk, and measures quality.
The implications reach far beyond data centers.
Here’s what stood out from our conversation.
1. The Transition from GC to Owner: Harder Than You Think
Tim was candid about the culture shock of moving from the contractor side to working on the owner side. As a superintendent, you’re used to calling the shots. Need the electrician to work on something? You tell the foreman directly. But on the owner side, you can’t just direct trades anymore. You work through the GC, through layers of management.
“Even though you are the owner, you don’t have the ability to just direct. You have to rationalize the why behind it and you have to have a justification.”
The learning curve was steep. Tim had to shift from command and control to leading through influence. Instead of providing solutions, he learned to ask leading questions that help GCs arrive at answers themselves. Rather than saying “we need to run two shifts,” he asks, “what opportunities do you see to compress the schedule?”
Perhaps that habit of collaboration that working through layers of management requires should be our default even when that constraint doesn’t exist.
2. Quality Standards Have Changed Dramatically
This was the most eye-opening part of the conversation. In the enterprise era, when Tim and I were building together, everything was platinum standard. Owners wanted the best reliability, the lowest downtime, perfection.
Then came cloud computing. Still high standards, but more cost and schedule pressure because external customers were involved and commitments had been made.
Now, in the AI and ML era? Everything has changed.
“The lost revenue of not turning over the building exceeds that of shooting for maybe a higher quality product. The risk tolerance overall on the owner side has greatly increased.”
Translation: speed to market matters more than perfection. Every day without capacity is lost revenue that exceeds the value of exhaustive testing. That doesn’t mean corners are being cut recklessly. It means owners are making more sophisticated risk-based decisions.
Tim gave the example of electrical acceptance testing. Instead of blindly following every test in the NETA standard, they now evaluate each component. How often do CTs actually fail? What’s the cost per CT to test versus the cost to replace if it fails? Is there redundancy in the design that reduces risk?
The traditional five-nine uptime mentality (99.999% availability) doesn’t apply to ML training. If training pauses for maintenance, users don’t see an error message. It happens quietly in the background. There’s still lost revenue, but the direct impact is different.
The industry is still figuring out the acceptable level of risk. But the old platinum standard is gone.
3. Labor and Equipment Are the Real Constraints
Steel and precast used to be the critical path items. Get those packages bought out first or the whole job stalls. Those supply chain issues still exist, but they’re not the primary bottleneck anymore.
Now it’s trade labor and equipment throughput.
The AI boom is exploding. Every hyperscaler is building as much and as fast as they can. But the labor force hasn’t grown at the same pace. And you can’t fast-track an electrician. Apprenticeships take four years.
On the equipment side, it’s not material availability anymore. Post-COVID supply chain issues have mostly resolved. The problem now is manufacturing throughput. How fast can Siemens, Eaton, and ABB produce switchgear? Their assembly lines face the same labor constraints as jobsites.
Add in the reality that server technology changes almost daily, and you have a recipe for constant obsolescence. By the time a building is delivered nine or ten months after groundbreaking, the server technology is already outdated.
4. Design Standardization Doesn’t Exist Yet
We asked Tim about standardization. Could you build a backbone infrastructure that allows for quick plug-and-play upgrades as technology evolves?
His answer was blunt: “Unfortunately, with how new AI and ML is, standardization just doesn’t exist.”
Power is easier to manage. Power is power. The only variable is how much you need. But cooling? That’s the real struggle. Different servers require different cooling strategies based on heat output. When server technology changes constantly, standardizing the cooling infrastructure becomes nearly impossible.
It’s a moving target. And until the technology stabilizes, every project is custom.
5. Culture Is the Differentiator
We asked Tim what separates average GCs from great ones. His answer had nothing to do with technical capability.
“The more successful GCs are the ones with frontline leaders that know and understand culture and how to build a positive culture on the site.”
In an environment defined by constant change, aggressive schedules, and high stress, the sites that deliver are the ones with strong culture. Frontline leaders who embed a growth mindset. Teams that are open to change and agile enough to adapt as intent shifts.
You can see it in productivity. The sites that build faster and meet impossible deadlines are the ones where there’s a positive culture on site. It’s not something you can easily define, but you know it when you see it.
Tim also emphasized the importance of understanding different trade cultures. Electricians operate differently than ironworkers. They have different mindsets, different worldviews. Strong leaders recognize that and adjust their approach based on the audience.
6. AI Is Automating Commissioning, But Humans Aren’t Going Anywhere
Tim is optimistic about AI’s role in commissioning. It has the potential to automate a lot of testing and make commissioning agents more efficient. But the human element isn’t disappearing.
“AI is only as good as the human with it. The output all depends on the command prompt from the person who put it in there.”
Commissioning agents won’t be replaced. They’ll be empowered. AI can handle repetitive tasks and surface insights faster than any manual process, but it still takes human judgment, experience, and empathy to act on those insights.
Various tools are in development, some construction-specific and some built in-house by tech companies trying to achieve their own capacity goals. But we’re still early. A lot is being tried, and the market hasn’t settled yet.
7. Advice for the Next Generation: Seek to Understand
Tim’s message to young project engineers grinding it out on remote jobsites was simple but powerful: seek to understand.
“Too often, you’re constantly being squeezed. But having the patience to seek to understand why something is the way it is will actually get you further.”
Don’t assume why something wasn’t done the way you wanted. Ask questions. Understand the situation. Then work together to solve it.
He also emphasized the importance of a growth mindset. Be open to alternative methods and strategies. Don’t be short-sighted. Yeah, maybe you don’t want to be in Amarillo right now. But if you stay two years and perform well, where could it take you?
Take a deep breath. Be patient. It’s a dynamic environment, but those who hang in there and develop that long-term perspective will go far.
Conclusion
This conversation with Tim Straub was a gut check for anyone who thinks data center construction is just about bigger budgets and faster schedules. It’s about recalibrating risk, adapting to constant change, and recognizing that culture, not technical prowess, often determines success.
The AI race isn’t slowing down. If anything, it’s accelerating. And the contractors who thrive won’t be the ones clinging to the old playbook. They’ll be the ones who embrace influence over authority, seek to understand before reacting, and build cultures strong enough to weather the chaos.
Tim’s career arc from superintendent to owner gives him a perspective that’s rare and valuable. And his advice to the next generation is worth repeating: seek to understand, think long-term, and don’t lose your patience.
Because in this race for compute, the ones who stay grounded will go the farthest.
Watch it here:
https://www.youtube.com/@ConstructionConversations
Best Quotes
“Even though you are the owner, you don’t have the ability to just direct. You have to rationalize the why behind it.”
“The lost revenue of not turning up the building exceeds that of shooting for maybe a higher quality product.”
“By the time you finish a building and you deliver it, it’s already out of date. The server technology has changed.”
“The more successful GCs are the ones with frontline leaders that know and understand culture and how to build a positive culture on the site.”
“AI is only as good as the human with it. The output all depends on the command prompt from the person who put it in there.”
“Seek to understand why something is the way it is. That will actually get you further.”
“Take a deep breath, be patient. Those that are patient and hang in there, I have an optimistic outlook that your career will go far.”
Originally published in Construction Briefs. Read on Substack →