Why TeraWulf Is Building Out Battery Storage!

Why TeraWulf Is Building Out Battery Storage for Its AI Factories: Inside the Company’s AI Overhaul and 2.3 GW Pipeline
TeraWulf (NASDAQ: WULF), a publicly traded Bitcoin mining company, has embarked on an ambitious transformation, positioning itself at the intersection of cryptocurrency mining and artificial intelligence infrastructure. In 2025, the company achieved a significant milestone by recording its first full year of revenue from AI operations, generating $16.9 million through the hosting of AI computing clusters at its Lake Mariner data center in New York.

Now, the company is setting its sights on a far larger vision. TeraWulf has identified five strategic sites that collectively support a planned 2.3 gigawatts (GW) of AI-focused capacity, marking a decisive pivot toward high-performance computing (HPC) and AI workloads.

Nazar Khan, TeraWulf’s Chief Operating Officer, recently appeared on the Blockspace podcast to provide an in-depth look at the company’s aggressive expansion strategy. According to Khan, the company is concentrating its development efforts on new sites in Kentucky and Maryland, where it is repurposing existing industrial infrastructure—a former aluminum smelting plant in Kentucky and a natural gas peaker plant in Maryland—to host advanced AI computing clusters. These new locations complement TeraWulf’s existing land and data center portfolio, which includes the active Lake Mariner campus in New York, a forthcoming site in Abernathy, Texas, and a proposed site called Cayuga, also in New York.

Kentucky Site: Repurposing a Former Smelting Plant for AI Workloads
One of TeraWulf’s cornerstone projects involves the redevelopment of a former Century Aluminum smelter site in Hodgenville, Kentucky. The location is particularly attractive because it previously supported a 480-megawatt (MW) power supply for industrial smelting operations, meaning that substantial power capacity is already available and ready to be redirected for AI data center use. The company is targeting 480 MW of capacity at this site by the second half of 2027.

Currently, cleanup and environmental remediation efforts are underway to prepare the site for the first phase of TeraWulf’s AI data center rollout. Khan emphasized that the site’s existing power infrastructure provides a significant advantage, enabling faster development timelines compared to greenfield projects.

A Strategic Advantage: Unconstrained Engineering
Khan highlighted a key differentiator for Bitcoin miners entering the AI infrastructure space: the absence of legacy constraints. Unlike traditional data center operators who may be bound by established methodologies and historical approaches, Khan noted that TeraWulf approaches each engineering challenge with a fresh perspective.

“We are not saying, ‘Hey, this is how we used to do it five years ago or three years ago or seven years ago,’” Khan explained during the podcast. “We’re coming out and saying, ‘What’s the issue? What’s the problem we’re resolving for, and how do we go about coming up with a solution for it?’” This mindset, he argued, allows the company to design more efficient, scalable, and purpose-built infrastructure tailored specifically to the demands of AI and HPC workloads.

Maryland Site: A Gigawatt-Scale Energy and Compute Campus
The second major development in TeraWulf’s pipeline is located at the former Morgantown coal facility in Maryland. This site originally housed a 1.5 GW coal-fired power plant, which was later converted into a natural gas peaker plant. TeraWulf now plans to transform this expansive property into a modern, integrated energy and compute campus.

One of the site’s most significant features is its robust power infrastructure. The location is served by nine separate transmission lines connected to three different utilities, providing exceptional redundancy and multiple sources of backup power. When fully operational, Khan explained, the campus is designed to function as a net supplier of energy back to the electrical grid—a distinctive characteristic that sets it apart from traditional data center developments.

“Our goal here really is to transform this into a campus that’s bringing an incremental gigawatt of efficient gas power generation, gigawatt of load, paired with 500 MW of battery storage,” Khan said. He further emphasized the strategic value of this configuration, noting that “these are all grid resources, both from a gen perspective and from a load perspective.”

The Role of Battery Storage: 500 MW for Grid Stability and Operational Efficiency
A critical component of the Maryland campus is the integration of 500 MW of battery storage, which is planned to be energized between 2028 and 2029. This battery system is designed to serve a dual purpose. First, it acts as a peak-saving mechanism for the broader electrical grid, helping to stabilize supply and demand fluctuations. Second, it provides a buffer for the variable power consumption patterns typical of AI and HPC workloads, allowing TeraWulf to manage energy costs and maintain operational continuity even during periods of high demand or grid instability.

Khan noted that the ambitious timeline for the Maryland project—spanning several years—requires close coordination with state and county officials. By aligning its infrastructure development with broader regional grid requirements, TeraWulf aims to ensure that the facility not only meets its own operational needs but also contributes positively to the local energy ecosystem.

Pioneering a Standardized 168 MW AI Infrastructure Design
TeraWulf has also been at the forefront of developing a highly optimized infrastructure model tailored specifically for AI and HPC applications. In partnership with Fluidstack and Google (NASDAQ: GOOGL), the company has pioneered a standardized design based on a 168 MW critical IT capacity configuration, divided across four distinct data halls. Each data hall contains 42 MW of net critical IT capacity.

This modular, repeatable design allows for efficient scaling and has been noted as a framework that other public companies, such as Cipher (NASDAQ: CIFR), are also beginning to adopt for their own site developments.

“It’s really the architecture of the hardware that’s using capacity,” Khan explained when discussing the layout. By standardizing the design, TeraWulf can streamline construction, reduce costs, and accelerate deployment timelines across multiple sites.

The Transition Away from Traditional Mining
As TeraWulf continues to expand its AI and HPC capabilities, Khan acknowledged that the company’s traditional Bitcoin mining operations will likely be cannibalized over time. The existing mining footprint is expected to be gradually repurposed or phased out as more AI capacity comes online, reflecting a broader industry trend in which cryptocurrency miners are repositioning themselves as critical infrastructure providers for the growing AI sector.

Through this multifaceted strategy—combining brownfield site repurposing, large-scale battery storage, standardized infrastructure design, and a commitment to grid participation—TeraWulf is positioning itself as a leading player in the rapidly converging worlds of digital asset mining and artificial intelligence computing.