You adjusted your strategy from a Bitcoin miner to what you are now in the space of two or three years. Did you ever expect this, and how does it feel to be at the other end of this revolution?
Respectfully, we were never just a Bitcoin miner. I've spent my entire career building and operating energy infrastructure. For more than 35 years, my team and I have developed, owned, and operated large-scale power assets. We've built more than six gigawatts of power infrastructure. That's the lens through which we've always viewed this business.
What changed over time was the best use of a megawatt. As power markets evolved, the long-term utility contracts that traditionally supported power projects became shorter and less financeable. Electricity increasingly became a spot market, and I don't like taking unnecessary risk. So, we started looking for ways to create more value from our power infrastructure.
Bitcoin mining was one answer. What attracted me wasn't Bitcoin itself, it was the flexibility of the load. You can turn it on and off almost instantly, which makes it a very useful tool for managing power assets. But I never liked the volatility or the difficulty of financing a business tied to a commodity price.
When high-performance computing and AI emerged, it felt like coming home. We took the same energy infrastructure expertise we'd spent decades building, developed a reference design for data centers, and paired it with long-term contracts backed by high-quality counterparties. That's a business I understand. Fifteen years from now, I know what those assets are expected to produce. For someone who has spent a career in infrastructure, that's a much more natural fit.
The best part is that I'm doing it with the same core team I've worked with for 18 to 25 years. We've been through multiple cycles together, and that experience matters.
Where do you see the bottleneck now? Is it permitting, the interconnection queues, something else?
Without question, the biggest bottleneck is power.
The demand coming from AI and data centers is extraordinary, and the country simply doesn't have enough power-ready infrastructure to meet it. But the answer is more complicated than just building more generation. Power plants don't move around. Every project sits within a specific regulatory environment, transmission system, and permitting framework.
You can build generation, but you also need the transmission infrastructure to move that power reliably and responsibly. That takes time, capital, and planning. Choosing the right location and jurisdiction is absolutely critical.
Permitting has always been a challenge. Nobody wakes up wanting a power plant or major industrial facility in their backyard. But I've found that transparency, education, and responsible development go a long way toward addressing those concerns.
One of our advantages is that we focus heavily on brownfield redevelopment. We can take sites that have already supported major industrial operations, like the former aluminum smelter in Kentucky, clean them up, repurpose existing infrastructure, and bring new economic activity to those communities. In many cases, that's much faster than starting from scratch.
Grid operators are also placing increasing value on projects that support overall system reliability. That's one of the reasons we're pursuing projects where power infrastructure and large-scale computing can be developed together.
Some people are looking at islanded or behind-the-meter generation. What's your view?
At least today, I'm skeptical.
The theory sounds good: you build your own generation, your own data center, and control everything yourself. But in practice, it often means building significantly more generation than you actually need just to maintain redundancy.
If you're building a 1,000-megawatt data center and need 1,500 megawatts of generation to support it, someone has to pay for that extra 500 megawatts sitting idle. That's a very expensive insurance policy.
You also have to think about what happens if the project doesn't perform as expected. If you're in a remote location without robust transmission infrastructure, how do you monetize that power? How do lenders get comfortable with the risk?
I still think most customers prefer a strong grid-connected model, drawing power from the grid while also contributing reliability and flexibility back to the system.
I'd also be cautious about very large projects located far from robust transmission infrastructure. AI workloads require extraordinary reliability. These aren't facilities that can tolerate frequent interruptions. The power system has to remain balanced and resilient.
That's one reason we emphasize geographic diversity. We don't want all our infrastructure in one location, and neither do our customers. As AI becomes increasingly important, the security and resilience of the underlying power supply become just as important as the data center itself.
Equipment is a concern for a lot of people. Gas turbines can take five or six years to get. Does that worry you?
Honestly, I think it's a little overblown.
We're power developers. We look at these issues differently than people who are entering the space for the first time. Yes, certain equipment can be difficult to procure, but there are also assets available from projects that have been delayed or cancelled.
My chief operating officer spent nearly two decades at Siemens. We've spent our careers solving these kinds of problems.
The real challenge isn't finding equipment, it's execution.
Operating a facility isn't what keeps me up at night. Building it is. At our Lake Mariner site, I had people hanging steel in 30-mile-per-hour winds off Lake Ontario in the middle of winter. Those men and women were heroes. That's what execution looks like.
A year from now, the winners in this industry won't be the companies that announced the biggest projects. They'll be the companies that delivered on time and on budget.
That's what customers care about, and it's what hyperscalers increasingly evaluate when deciding who to work with.
Of all your competitors, which do you admire?
I don't spend much time focusing on competitors, and I certainly don't want to criticize anyone.
CoreWeave has built something impressive and has clearly reached a different stage in its development. Nebius has assembled a very smart team and built a strong cloud platform.
Cipher strikes me as a thoughtful and disciplined organization. Hut 8 has executed some very good transactions.
But ultimately, this business comes down to execution. Announcing a deal is one thing. Delivering it is something else entirely.
What differentiates TeraWulf is our team. The top 25 people in my company have worked together, on average, for roughly 15 years. We've seen markets go up and down. We've built projects in difficult environments. We've worked through problems together.
Customers don't care about corporate talking points. They care about transparency, reliability, and execution. That's what we focus on every day.
You've had an incredible run, around 500 percent in a year. Does the equity still have room to run?
We've told investors exactly what we're going to do: deliver between 250 and 500 megawatts of new contracted IT capacity every year.
At roughly $10 million per megawatt, that's a tremendous amount of infrastructure to build. Some people have said they'll do more, and that's fine. But I graduated from Annapolis, and in the submarine community we had a saying: one "aw shit" wipes out ten "attaboys."
I'd rather be conservative and deliver every commitment we make than overpromise and disappoint.
If we continue executing, continue signing long-term contracts with high-quality counterparties, and continue delivering projects on schedule, the market will ultimately determine the value of that performance.
I remain highly aligned with shareholders. I'm the company's largest shareholder, and our management team collectively owns a significant stake in the business. The board has been together since our transition to the public markets, which allows us to move quickly and stay focused on long-term value creation.
The attention, the capital, and the excitement around AI haven't changed how we operate. We still do the same things we've always done—investment committee every Monday, operations review every Thursday, and relentless focus on execution.
