In the second part of Kevin Brown’s conversation with Nazar Khan, Co-Founder and Chief Technology Officer at TeraWulf, they explore how the company has scaled its Lake Mariner campus from a two-megawatt pilot project into one of North America’s largest AI and high-performance computing developments.

Nazar explains why TeraWulf chose to invest in infrastructure before securing its first customer, beginning with a small proof of concept before progressing to increasingly larger data centre buildings. By demonstrating that it could design, construct and operate the infrastructure in practice, the company was able to build confidence among customers, lenders, technology providers and other partners.

The conversation also examines what it takes to deliver AI infrastructure at speed. From securing billions of dollars in financing and forecasting equipment requirements to coordinating utilities, suppliers, contractors, customers and local authorities, Nazar highlights the importance of collaboration across the entire ecosystem. He also reflects on Lake Mariner’s impact on the surrounding community, transforming a former coal-fired power station into a growing digital infrastructure campus supporting hundreds of permanent roles and approximately 1,600 construction workers.

Featured guest: Nazar Khan, Co-Founder and Chief Technology Officer, TeraWulf

Key quote: “The best way to convince someone that you can actually get it done is to go get it done.”

Key takeaways:

• TeraWulf began with a two-megawatt pilot project to develop the skills, designs and partnerships needed to build AI infrastructure.

• Investing in infrastructure before securing a customer helped the company demonstrate its capabilities and build market confidence.

• Access to capital is essential, as AI-ready data centres can require investment of approximately $9 million to $10 million per megawatt before IT hardware is installed.

• Delivering infrastructure quickly requires close coordination between customers, utilities, finance teams, equipment providers, contractors and local authorities.

• Flexible design decisions and a willingness to evaluate trade-offs can help projects overcome equipment availability and supply-chain constraints.

• The redevelopment of Lake Mariner shows how former industrial and power generation sites can support new employment and become part of the AI economy.

Follow or subscribe to The Machine Room on your preferred podcast platform for more conversations with the people building the power, cooling and digital infrastructure behind AI.