Weekend Interview: Iceotope’s Nathan Blom On AI’s Future And Building Greener Data Centers
This series goes deep with some of the most compelling figures in commercial real estate: the dealmakers, the game-changers, the city-shapers and the larger-than-life personalities who keep CRE interesting.
Nathan Blom’s career has taken him from the pulpit to the C-suite.
Blom was named this week as one of two interim CEO’s of Iceotope Technologies, a firm specializing in advanced cooling systems critical for data centers that support artificial intelligence.
But the Denverite’s career path didn’t begin in data centers, tech or IT. His journey began in the church, pursuing a career as a Lutheran pastor that saw him lead congregations in three states before deciding to shift his focus toward the business world.
At Iceotope, Blom has served as chief commercial officer for nearly two years following extended stints at HP and Lenovo. He now takes the helm as the firm positions itself to play a central role in AI’s transformation of the data center landscape.
AI is driving an unprecedented wave of data center development, but it is also changing how these facilities are designed and built. Data centers have traditionally used circulated air to cool the computing equipment within them, but the high-performance computing equipment for AI gets far hotter than the servers older data centers were designed around.
Because of AI, new data centers are being designed to support liquid cooling — systems that use liquid refrigerant to remove heat directly from the IT equipment itself. Iceotope designs and produces sophisticated liquid cooling systems that pump fluid inside data center IT equipment to cool specific components directly.
While leading a firm focused on enabling the development of artificial intelligence may seem like a far cry from his earlier clerical roles, Blom says his work at Iceotope feels similarly purpose-driven, particularly due to the technology’s potential to reduce data centers’ environmental impact.
Blom sat down with Bisnow to discuss whether or not he’s worried about an AI bubble, helping the data center industry break old habits and the lessons he learned as clergy that he’s carried to the board room.
This interview has been edited for length and clarity.
Bisnow: You're about to step into the interim CEO role at Iceotope, but your career began on a very different trajectory as a pastor. How did you end up going from the clergy to your current career path?
Blom: I found a calling in the church early on in my life and decided to pursue full-time ministry as a career directly out of college. I went to seminary and served at congregations in Minnesota, Colorado and Texas. These were big congregations, and there were executives who were mentoring me since I was still new to my career. This inspired me to learn more about how to operate an organization.
With their encouragement, I got my MBA at the University of Texas at Austin while I was still working in the church. That sparked a fire in me to apply my skills in new ways that would still help people and have an impact on the world. That’s translated to my role at Iceotope. We’re doing something that has a material impact on our environment and our climate, and that gives me a reason to get up in the morning and stay up late at night thinking about this stuff.
Bisnow: So, they’re both purpose-driven careers. What skills or lessons from your work in the church do you find yourself applying in your current role?
Blom: The number one thing is caring about people — being empathic to your customers' needs and listening. It’s also supporting your staff and other people around you who are trying to be successful and help you be successful at your task.
The other thing is having a strong culture and mission, and knowing why you're doing what you're doing so that you don't get distracted. Whether in the church or in business, there are thousands of good things you could pursue but you have to choose only a couple of them. Having a very strong mission and knowing exactly why you exist as an organization really matters.
Bisnow: Let’s talk about Iceotope. You guys are one of the beneficiaries of the trillion-dollar wave of AI infrastructure spending that's underway. AI is changing how data centers are designed and built, particularly when it comes to cooling. What's Iceotope’s role in this transition?
Blom: We’re at an inflection point where AI and other technology is making the equipment in data centers hotter than in the past. Think about Nvidia, the company making the GPU’s that power a lot of AI. Those things are getting hotter and hotter, and they have now exceeded the physical capacity of air to cool them, which is how it has typically been done until now.
That's where Iceotope steps in. We’ve developed a patented technology we call precision liquid cooling. We use a fluid that does not conduct electricity, which is a much better way of capturing heat than air. The fluid becomes highly efficient at grabbing the heat from these really hot components so that we can continue to expand the compute power that AI demands.
We do this using the least amount of electricity and the least amount of water that we can. This is important because not only are consumers worried about the environment, but governments are getting tired of how much electricity data centers consume and the stress that’s putting on the grid. We need to make sure that as many electrons as possible go toward the actual computing and not to ancillary functions like cooling.
Bisnow: There's a lot of different options out there when it comes to liquid cooling in data centers: refrigerated plates that sit against servers, baths of dielectric fluid that IT equipment is submerged in. What's unique about Iceotope?
Blom: We deliver fluid individually to every component within the ecosystem. So, the power supply, the memory, the SSDs, the CPUs and GPUs all received their individual allotment of dielectric fluid to capture 100% of the heat. And so that means that we're able to be more efficient.
And we do that in a format that’s the same as in a standard data center environment. That’s important because it maintains the common form factors that we're used to, meaning you can avoid taking up a much larger footprint and you can maximize the use of the data center building you already built. This has massive environmental impacts in terms of the use of concrete and other resources to expand or build new buildings.
Bisnow: Adoption of liquid cooling in data centers can be viewed as an indicator of corporate AI adoption more broadly. You have your boots on the ground here. Where are we on the demand curve? You used the term "inflection point," but are you seeing a spike in orders or is it just conversations at this point?
Blom: We’re right in the middle of that spectrum, honestly. In order for liquid cooling to be adopted at scale, data centers need to have the infrastructure in place to enable it. So, the leading indicators of liquid cooling adoption are the revenues of companies like Schneider Electric and Vertiv that produce the systems that need to go into the ecosystem first. If you look at the order books for the data center infrastructure side, their backlog is dramatic right now on liquid cooling. They're trying desperately to ramp up production to meet the almost unlimited demand they're seeing. They're getting billion-dollar individual orders from the big tech companies.
On the colocation side, a lot of their tenants are doing research right now on which technology is best for them. What’s going to happen is that they’ll suddenly tell the colocation provider the cooling solution they want, and they’ll want it immediately. If the data center provider isn’t ready and needs 12 months to get the facility ready, that doesn't bode well for their future.
From a real estate perspective, from the discussions I'm having with the colocation data center firms and with the major hyperscalers, they have no intention of building a non-liquid-cooled enabled data center going forward. This is already happening.
Bisnow: What about older data centers? There are mixed opinions on whether there’s a good business case for retrofitting legacy data centers with liquid cooling so they can accommodate AI. Do you think there will be a market for AI retrofits?
Blom: I think there will be. The hands of legacy data center operators are going to be forced. Any operator that wants to play in this part of the AI market will have to at least some kind of retrofitting. At a certain point, it’s mandatory to use liquid cooling for performance reasons. There's no way to not use a liquid-cooled environment to do those kinds of things.
There will continue to be air-cooled data centers, there’s no question about that. If you have a lower-end workload it doesn't necessarily make economic sense to spend on liquid cooling. But there’s definitely a market there for retrofits. Companies that want to remain relevant are probably going to have to make the leap at some point.
Bisnow: You’ve brought up sustainability a number of times. There is growing public discussion about data centers’ significant power and water use, particularly surrounding AI. Can you mitigate some of these environmental impacts?
Blom: Just to start with some basic numbers — we will use 40% less energy and we will use nearly 100% less water. That either frees up electricity for the operator of the data center to redeploy towards compute that’s actually profitable, or it saves the grid from having to deliver that 40%. That’s a powerful impact right up front.
And when you're looking at construction costs, it turns out that the capital expense isn’t prohibitively different than if you were doing a traditional air handling system. So, from an apples-to-apples comparison for a new build, it makes sense.
Bisnow: So, at this transitional moment for the data center industry, what’s the biggest challenge facing Iceotope?
Blom: Our number one challenge is the status quo. There’s a strong gravitational pull to continue doing things the way they’ve always been done. AI is causing a revolution — and that’s really what it is — but behind the scenes, it’s also dramatically changing what the future of hardware architecture is going to look like. There’s a lot of hesitation to change with enterprises, so our greatest challenge is making this move to an alternative technology feel relatively safe and easy.
Candidly, in the last six months we've started to see a very different tenor of conversation than before. People are recognizing that it's not an option anymore. This isn't a theoretical exercise. It's something that they need to start acting on.
Bisnow: There’s been a lot of talk in recent weeks about an “AI bubble,” and there are questions about how much of the data center buildout happening right now is speculative. From where you sit, smack dab in the middle of the AI value chain, how do you view the potential risks on the demand side? And how do you prepare for the possibility that AI adoption falls short of expectations?
Blom: Of course there's speculative investments happening. That's the nature of the beast. But a lot of the bubble talk seems to be emanating from Wall Street investors who are looking to the big technology firms and saying, “You've invested billions of dollars in AI a quarter ago, where's the return?” Because if you’re a Wall Street investor, quarterly returns are paramount. The reality is that those large technology firms are investing for the long term. They're not investing for immediate ROI.
For us as a company, our strategy aligns with those long-term investments. We have a powerful IP, and we'll continue to apply that IP to each generation of technology as it emerges. We look at this from a longer arc of history. We're not worried about what's going to happen in Q3 of this year so much as we’re thinking about Q3 of next year and Q3 the year beyond.
Bisnow: Can you give us your bold prediction for this year?
Blom: AI is not going anywhere. There may be a bubble, but the burst is going to hurt the people who didn't understand the technology to begin with and just got on the AI bandwagon. There’s plenty underneath that bubble that's going to continue.
You’ll see the same thing that happened with the cloud industry, which consolidated into a handful of very large players. AI will have a small number of very large players who are going to make AI relevant for things we haven't even imagined yet.
Bisnow: This is a weekend interview, so what’s your favorite weekend routine or activity?
Blom: I live in a great place to be outdoors here in Denver, so it's going to sports events, soccer games or baseball games, or stuff with the kids. It’s also spending time enjoying the mountains.
