Q&A With Steve Hammond: AI Data Center Challenges and NLR Solutions
How Steve Hammond Grew the Laboratory’s Computing Roots With Bold Ideas and Colleagues
This article is part of the Manufacturing Masterminds profile series, which provides an inside look into the lives, research, and impact of NLR's advanced manufacturing researchers.
Steve Hammond’s career path has been charged by curiosity and collaborators.
He worked on powerful turbofan engines and delicate light bulb filaments. He helped improve the accuracy of MRIs and ensure that stealth aircraft remain evasive. He knows what it feels like to be a foreign national overseas as a postdoctoral researcher and what it felt like to watch the Twin Towers fall on Sept. 11, 2001.
“When 9/11 hit, I suddenly felt I needed to make a change,” Hammond said. “I wanted to do something to address our problems directly.”
That urge brought him to the National Laboratory of the Rockies (NLR). There, in 2002, he built the laboratory’s Computational Science Center from the ground up and became its first center director. And he grew the center to 100 people during his time in that role. “Those are my roots,” he said.
And now after almost 24 years at NLR, his (grey hair) roots are his badge of honor.
“I’ve seen a lot of changes, and I’ve got a few grey hairs to show for it. But I’ve had a lot of fun,” Hammond said. “I’ve worked on problems that I'm passionate about with people who I enjoy working with—and that makes it worthwhile.”
In the latest Manufacturing Masterminds Q&A, Hammond explains some of the big artificial intelligence (AI) data center challenges, why NLR is the laboratory to solve them, and the qualities we need in our next generation of energy problem solvers. This interview has been edited for clarity and length.
When you were a kid, did you imagine yourself becoming a scientist?
I thought I'd be a lawyer because I argued a lot. In college I majored in physics and planned on becoming a physicist. But I saw some brilliant physicists waiting forever for an academic faculty position to open. So, I switched majors to math and computer science and then got a Ph.D. in computer science. That was the beginning for me: learning to be flexible, not necessarily having your mind set. And seeing the “white space” in a problem as an opportunity to explore what’s possible.
Did you find any “white space” when you were building the lab’s Computational Science Center from scratch?
First and foremost, the people we hired were just incredible. The lab wasn’t known as a leader in computing, but they were inspired by the lab’s mission and a shared vision to make a difference. Next was the opportunity to build not just a new data center but the world’s most energy-efficient one.
Ben Kroposki and I were technical leads to develop the Energy Systems Integration Facility (ESIF). Ben focused on the high-bay labs and power systems areas. I focused on the data center portion. For the data center, we had specific efficiency targets: not just something to meet our existing and future computing needs but a dual mission to also be a place where we could advance data center technologies, too.
I’m guessing you didn’t have many existing examples to model after and had to make some bold choices?
The audacity to do what we did—to go with liquid cooling and to pick Hewlett Packard to partner with us—we took on an incredible challenge and we changed the industry. We were the first to do component-level, warm-water cooling. We were the first to have a high-performance-computing data center that didn't have chillers; we used evaporative cooling instead. We were the first to capture the heat from the data center and use it effectively. That affected the design of the building and the integration of the thermal system. It was just sort of unheard of. And to this day, we still capture heat from the supercomputer for snowmelt in the plaza area around the ESIF. We helped Hewlett Packard launch their Apollo line of liquid-cooled high-performance-computing systems and were recognized with an R&D 100 award for our efforts.
Our thermosyphon “first” also involved partners who were able to demonstrate and deploy their tech here at the lab, right?
Yes. Johnson Controls reached out to us and said, “Hey, I've read about what you do in your data center. We have some ideas about how to reduce your water usage.” So, we partnered with Dave Martinez and colleagues at Sandia National Laboratories and Johnson Controls to redo the plumbing in the data center. Johnson Controls demonstrated their prototype thermosyphon system, and when we installed it beside the evaporative towers on the ESIF roof, it cut our water usage in half without sacrificing efficiency.
Are we still pursuing the boundaries of what’s possible at the lab today?
Absolutely. The ESIF is still a living laboratory. We can help industry demonstrate new technologies as part of an integrated data center environment. We know what types of data center technologies are necessary for AI data centers—and we've been doing this for over a decade.
What kinds of challenges are we facing in the United States when it comes to data center demands and AI?
There’s no single technology that's going to ensure AI leadership. The chips are coming along. But as you start to build from chips to racks to systems, we’re headed toward unprecedented power density. When I started at the lab, we thought it was incredible to imagine going up to 50 to 60 kilowatts in a rack. Today, Nvidia's talking about one megawatt in a rack by 2027. That's a massive increase in power density contained in the same footprint of your household refrigerator. What's the right voltage? How do you do it safely? How do you effectively get the heat out and then reject it from the facility? How do we reduce the water needed and effectively generate enough power needed by these massive facilities? There are tremendous challenges being faced.
And beyond the extra power density and heat to account for, you’ve got a massive operational challenge, too.
Absolutely. Enterprise computing is email and web surfing and cat videos, where small bits of processing all work independently, and on average you have a pretty even load. AI data centers are fundamentally different in that you're training large language models with almost every processor working together. So, you go from very low power to nearly full power in a few clock cycles, and that can be disruptive to the grid. We need to identify and demonstrate the right infrastructure needed to manage and mitigate those transients.
That’s another big question: How will data centers for AI fit into the grid?
You need systems integration and coordination. Think about an airport with no air traffic control and all the planes trying to coordinate takeoff and landing on their own. You need some equivalent of air traffic control for the grid to coordinate the behavior of all these data centers and to ensure that the grid is reliable and affordable.
Can NLR help find that air-traffic-control-type solution?
Yes, we can reduce risk by ensuring that the necessary performance is demonstrated in these integrated systems before they’re deployed. You need to demonstrate the capabilities and performance of individual technologies, but you also need to do integrated piloting at a data-center scale to show that they all work together.
We can do that at NLR. Industry and lab partners have used our Advanced Research on Integrated Energy Systems (ARIES) to get a grid perspective—everything from power electronics to the controllable grid interface and all the infrastructure—to evaluate and understand data center behaviors and how to control them. Our focus on data centers also complements the U.S. Department of Energy’s Genesis Mission so that we have both the AI software and infrastructure in place to help ensure U.S. leadership.
When you think about the next generation of problem solvers who are wondering whether they can succeed in an energy-related career, what would you say to them?
One: You need to be flexible. Opportunities will present themselves, so be ready to embrace them. Two: Bring your best ideas, because I think everybody's got a superpower. And three: They say to find something you love and call it work. If you can find that thing that aligns what you enjoy with your vocation, it’s not a job anymore. It's just fun.
Read other Q&As from NLR researchers in advanced manufacturing, and browse open positions to see what it is like to work at NLR.
Last Updated April 28, 2026