Community, Compute, and the AI Race: Nancy Coblenz on Building the Right Infrastructure
Insights
- Most enterprises approach data center infrastructure as a utility decision. The organizations getting it right are treating it as a community and innovation platform.
- Buildings are being completed with no power to connect them, and power is being committed with no buildings ready to receive it. The sequencing problem in AI infrastructure is real and worsening.
- The next competitive frontier isn't just who builds the most compute. It's who builds it with enough intentionality to earn the policy, community, and regulatory trust to keep it running.
AI infrastructure is moving faster than the systems built to support it. Nancy Coblenz, President and co-founder of Stellenium, sat down with Jeff Kavanaugh, Head of the Infosys Knowledge Institute, at the Semafor World Economy Summit in Washington, D.C. to discuss what a more deliberate approach to data center development looks like in practice. Where most builders are optimizing for speed, Stellenium works backwards from use case and community impact, a model Coblenz argues unlocks better long-term outcomes for enterprises, municipalities, and the innovation ecosystems around them. The conversation covers the compounding mismatches between buildings, chips, and power availability, the policy backlash beginning to emerge at the state level, and why the US cannot afford to treat AI infrastructure as an afterthought in a global race where others are not slowing down.
Jeff Kavanaugh:
Welcome to a series of conversations with leaders who are turning AI's promise into real business results. I'm Jeff Kavanaugh, and today we're joined by Nancy Coblenz, President of Stellenium. We're here at the Semafor World Economy Summit in Washington, D.C. Nancy, welcome.
Nancy Coblenz:
Thank you so much, Jeff. I'm so excited to be here.
Jeff Kavanaugh:
I am too. Data centers, power, exciting stuff.
Nancy Coblenz:
Yes.
Jeff Kavanaugh:
What is the most surprising impact from your perspective of the AI revolution and how will it redefine how companies compete?
Nancy Coblenz:
The most surprising thing is the sheer amount of demand when it comes to compute power. As a human race, we've never consumed less power and less compute power. We will always consume more. No matter how efficient we get, we'll find new ways to use more. So we are just supporting that ecosystem and how on the infrastructure side we can build out to that.
Jeff Kavanaugh:
How is enterprise AI reshaping a company's external competitiveness while it's also working on the productivity aspects?
Nancy Coblenz:
Having a culture of internal innovation and entrepreneurship is so key in being able to have every single person be a steward of that innovation within the company. It really trickles down, whether it's an internal AI tool that we're building out and that bleeds into the engineers who have to create the software layer down to the compute power. Then the exact opposite is how we actually take our own solutions, whether it's a hard product or software, and provide the right solutions for our own clients and customers. There's a balance between the two, and I would say the ones that are doing the best at it understand that you need to have that entrepreneurial spirit within the company itself to really make it. It's not just cogs and wheels anymore.
Jeff Kavanaugh:
Going a little deeper in that, you're a data center company. On the one hand, you could say you're just a periphery, you're just plugging in. But it's more pervasive than that. How does what you do influence that culture and a company's ability to implement their strategy?
Nancy Coblenz:
We pride ourselves on being a technology-first organization as opposed to infrastructure-first. So it's a different positioning and modality that we put into designing the data centers and how we include the power supply into the type of data centers. There are a million different ways of actually utilizing a building, putting racks in it, putting GPUs and servers in it. It's all about the actual use case. What are you going to use it for? What is the actual impact it's going to have on your community?
Jeff Kavanaugh:
And we spoke and you said depending on whether it could be a hyperscaler or enterprise, they have all these different uses for it.
Nancy Coblenz:
Even governments. The typical model right now is, we will build it, they will come. We take a more pragmatic, intentional lane where we build on the compute layer first and say, what do you actually need to use this for, and then build backwards. That has helped us unlock the next generation of where things are going. A lot of the technologies that we are integrating, whether it's on the cloud compute, the data center infrastructure itself, or the power plants, we are waiting for allocation because they haven't been built yet. That's the crazy part.
Jeff Kavanaugh:
So you've got a building with no chips and people like Microsoft have chips and no buildings.
Nancy Coblenz:
Yes, but now the biggest issue is that there's no power. Buildings are going to be built and then there's no power to plug in. So then you have a building just sitting there for two to three years because there's no power.
Jeff Kavanaugh:
Well, at least the power industry has gotten very relevant, haven't they?
Nancy Coblenz:
I know, right? So those are the kinds of gaps we have to solve for. It really is links in a chain that you have to figure out collectively together. It has taken two years of our team to collaborate, find the right innovators, the right producers, and the right manufacturing companies that can actually deploy. That's where the market really is going.
Jeff Kavanaugh:
With such significant uncertainty and rapid change, it requires adaptation. How have the operating models of the people you're working with changed?
Nancy Coblenz:
We do a lot of thought leadership and workshops internally to really prepare the teams and the workforce, to help them understand what they're getting into and how this is going to unlock so much more for them.
Jeff Kavanaugh:
So they don't really understand what they're getting into when they first contact you. You're educating while you're selling or helping.
Nancy Coblenz:
Yes. On the enterprise level, as an off-taker, they're really just looking for compute power and that's basically it: I just need to plug in and I just need to use.
Jeff Kavanaugh:
That's their initial view.
Nancy Coblenz:
That's their initial view. Then when we peel back the layer and help them understand that they have a more vital role within the community itself, to be able to support the community around them, that's something very unique to us. We're community-first in that way. This infrastructure impacts the people that are locally there. We have an opportunity for edge compute in the communities we're in now. So as opposed to just providing space for what the technology needs are, let's turn this into a community-first initiative and support the innovation ecosystem around there too.
Jeff Kavanaugh:
So you're taking data centers from something people may see negatively and turning them into a positive, helping innovation that wasn't going to be there by giving communities a tool.
Nancy Coblenz:
Absolutely. We have a program that we put all of this through. So that's the workshops, thought leadership, and direct municipal and county leadership engagement. We have them in the driver's seat with us to ensure that their community is not only heard and seen, but actually proactive in it too. That's been a game changer.
Nancy Coblenz:
The people that are building data centers currently today are doing what they need to do to just get their product out there. Rightfully so, that is their way of doing it. Because we're able to focus more on the two- to three-year buildout and online timeline for us, it gives us enough time to really dig deep with the communities we are working with, and then this bleeds into state legislature. Maine is trying to propose a moratorium of no data centers at all. There's a lot of information and a lot of confusion out there. At the end of the day, if we don't lean into and invest in our AI capabilities as a nation, others are going to do it. We really are in an AI race.
Jeff Kavanaugh:
Also following down that path, automation and AI have caused leaders to question their current talent models and workforce. I know you don't directly deal with that, but what changes do you see as they adopt more data centers and this new way of looking at it to optimize the human-plus-AI connection and keep humans more at the center?
Nancy Coblenz:
For us, we see it as really enabling and unlocking the innovation ecosystem.
Jeff Kavanaugh:
Beyond utility, it's actually this innovation. So it goes from this dry idea of utility and plugging in to something that's exciting and inspiring.
Nancy Coblenz:
Absolutely. It is a foundational platform for literally anything you can build and want to dream of. That goes into healthcare, education, the community, biosciences. The biggest part of what we are trying to power is the next generation of innovators. We can't do that without the AI compute to help them get there.
Jeff Kavanaugh:
Have you seen success in helping people understand this really unleashes their potential? Do you ever see the light bulb go off for them?
Nancy Coblenz:
Absolutely. Once we start peeling back the layers and mapping the pieces out as to what this unlocks for their own capabilities, they begin to see that there are certain skill sets they don't need to already be the expert in. I don't need to have 10 or 20 years of expertise in coding anymore. It unlocks an entire new world of capabilities for me to build something like a new cancer detection model.
Jeff Kavanaugh:
At the risk of a cliché, it democratizes the innovation process.
Nancy Coblenz:
Exactly. That comes from my background and my co-founder's background. We are both entrepreneurial exited founders who had to do it the hard way. We had to figure out how to do all this stuff ourselves. He's a software engineer and infrastructure engineer for the last 25 years. He built his own companies literally from the ground up, having to physically code for hours and hours and hours. We don't need to do that now. What it does is unlock us to be more creative and able to think of the most outlandish solutions out there and actually make them happen.
Nancy Coblenz:
My belief in all of that comes from my dad. He's the last living lead engineer from the Man on the Moon mission, and he's still alive at 96 years old. Growing up, he would tell me stories about how he had to invent new mathematical calculations to get astronauts to the moon. If I can do that, Nancy, you can do anything too. I really believe with what we're doing on the infrastructure layer all the way to cloud compute, we're unlocking that ability for other people.
Jeff Kavanaugh:
Well, it's amazing, especially coming on the heels of Artemis II. Isn't that amazing?
Nancy Coblenz:
Yes. It's more than amazing. I watched it with him and he cried. He was so happy seeing that.
Jeff Kavanaugh:
Absolutely. It's good to see that arc. It's also inspiring because whatever we've done, there's so much more we can do now. But the key is for more people to believe that.
Nancy Coblenz:
Exactly. Realistically, this unlocks accessibility for a lot of people that probably thought they couldn't be entrepreneurial or technical.
Jeff Kavanaugh:
If you talk about inequality, this is also a way to attack that because you're leveling the playing field.
Nancy Coblenz:
Totally.
Jeff Kavanaugh:
For $20 a month, or maybe it's subsidized, you now have access to that. There's no elite system somewhere. You can have that same access.
Nancy Coblenz:
Exactly. That's really been the topic that has helped us work in other countries as opposed to the U.S. In the U.S. there is already a framework for how we do infrastructure. But when we are talking to presidents of different countries or their ministers of technology, it's a different conversation where they know they need this technology, they just don't know what it unlocks and why. We fully provide an ecosystem, including knowledge transfer. If they want access to our best and brightest PhDs when it comes to AI compute, we have our own professors who can do residency programs with their universities. These are things they would have never imagined before.
Jeff Kavanaugh:
You mentioned PE. This is some kind of venture or privately backed. What is your plan? Are you going to take over the world with this? Are you going to grow quite a bit and plateau? Where do you think you're going to go with this? Because you're pretty early in the process.
Nancy Coblenz:
Very early in the process, yes. We want to obviously grow as far as we can, but we want to be intentional with the projects we take on because it is not just one element of the infrastructure spectrum, it's all of it. We are very specific on who and where we take projects on. But we're also trying to find a way to make it more scalable without losing the inherent mission and purpose that we have. That's always the balance.
Jeff Kavanaugh:
Innovation and scalability without losing one or the other.
Nancy Coblenz:
Exactly. Realistically, what my co-founder is building on the compute side is basically trying to eliminate infrastructure and the actual data centers themselves so we can have local compute power on our devices. That is where we're trying to get to.
Jeff Kavanaugh:
So the ultimate democratization of compute.
Nancy Coblenz:
Yes, but that's years out.
Jeff Kavanaugh:
Will they be up there floating around the sky with the Starlinks?
Nancy Coblenz:
Possibly. But also quantum. Quantum is always seven years out. Whenever quantum actually happens, that's the game changer that we don't know until we get to that point.
Jeff Kavanaugh:
But at least you're going to be ready with this established.
Nancy Coblenz:
Exactly. For me, what's most exciting is that it's providing a forum for people to be able to lean into something as boring as data centers and power plants and now see the beauty in what it can actually unlock.
Jeff Kavanaugh:
This is exciting. I think you provided a whole new perspective on the world of data centers.
Jeff Kavanaugh:
Nancy, thank you for your time.
Nancy Coblenz:
Thank you so much, Jeff. I really appreciate it.
Jeff Kavanaugh:
I'm Jeff Kavanaugh from the Infosys Knowledge Institute. Until next time, keep learning and keep sharing.