TeraWulf’s transition from bitcoin miner to AI infrastructure operator crossed an important milestone in the first quarter of 2026, as revenue from its high-performance computing (HPC) hosting business surpassed income generated from mining cryptocurrency for the first time. The company reported $21 million in HPC lease revenue during the quarter, compared with roughly $13 million from bitcoin mining, according to its first-quarter earnings release. Total revenue came in at approximately $34 million. For years, companies like TeraWulf built out massive energy-intensive facilities to mine bitcoin. Now, many of those same operators are finding that artificial intelligence companies are willing to pay far more for access to power, cooling systems, and ready-built data center space. “This is the first period where HPC leasing is meaningfully reflected in our financials,” Chief Executive Paul Prager said during the company’s earnings call. The change happening at TeraWulf mirrors a wider shift across the crypto mining industry. As mining profits fluctuate with bitcoin prices and energy costs, operators with access to large power supplies are increasingly repositioning themselves as infrastructure providers for AI firms and cloud computing customers. That pivot comes as the race to secure electricity for AI systems intensifies globally. The International Energy Agency said electricity consumption from data centers worldwide is expected to nearly double to around 945 terawatt-hours by 2030, with AI emerging as the main driver behind the increase. The agency also said data centers could account for nearly half of U.S. electricity demand growth by the end of the decade. From volatile mining revenue to long-term AI contracts Although TeraWulf’s overall revenue was little changed from a year earlier, the business itself is starting to look very different. AI hosting has quickly become TeraWulf’s biggest source of revenue, a departure from the company’s traditional dependence on bitcoin mining. Unlike mining income, hosting contracts tied to AI infrastructure tend to provide steadier and more predictable cash flow. Chief Financial Officer Patrick Fleury described the company as “a business in transition”. He believes that revenue is increasingly tied to “stable, contracted” compute agreements, according to The Block. TeraWulf reported a quarterly net loss of $427.6 million, compared with a loss of $61.4 million during the same period a year ago after investing in data center expansion and AI-related infrastructure. So, the transition proved expensive. Operating expenses rose to nearly $200 million during the quarter, partly due to impairment charges associated with scaling back segments of its bitcoin mining operations. Despite the mounting costs, investor enthusiasm around AI infrastructure companies has remained strong. However, companies with large amounts of power capacity and ready-built facilities are the potential winners in the race to support growing AI demand. Barron’s reported that TeraWulf shares have more than doubled this year as investors. So, it comes as no surprise that the company intends to focus on AI hosting ambitions rather than its legacy mining operations. “As we continue to scale, we expect the business to be increasingly driven by recurring, contracted revenue, reducing exposure to the volatility historically associated with bitcoin mining,” Fleury said in the company’s preliminary quarterly filing. Morgan Stanley analysts Stephen Byrd and James Faucette said they are seeing “increasing willingness among key AI players to pay higher ‘time to power’ premia in the form of increasingly rich economics to Bitcoin companies,” according to Barron’s. Expanding beyond bitcoin mining TeraWulf said it ended the quarter with 60 megawatts of operational HPC capacity at its Lake Mariner facility in New York, where AI cloud company Core42 is among its customers under a long-term lease arrangement. The company is also developing additional facilities at the site while coordinating deployments with customers, including Fluidstack and Google. Outside New York, TeraWulf has expanded aggressively into new power markets. Its Hawesville, Kentucky, project carries roughly 480 MW of grid-connected power capacity, while a Maryland site could eventually scale to as much as 1 gigawatt if regulators approve expansion plans. The company has previously disclosed more than $12.8 billion in long-term AI and HPC contracts tied to 522 MW of critical IT capacity. TeraWulf reiterated plans to add between 250 MW and 500 MW of new contracted capacity annually and said it held approximately $3.1 billion in cash and restricted cash at the end of the quarter. AI demand is reshaping the power market The scramble for AI infrastructure is beginning to alter the economics of electricity markets in the United States and beyond. Building entirely new data centers and securing grid access can take years, especially in regions already facing transmission bottlenecks. Former bitcoin mining sites offer a shortcut because many already have high-capacity power connections in place. The International Energy Agency said electricity generation dedicated to data centers could rise from about 460 TWh in 2024 to more than 1,000 TWh by 2030. Utilities are already revising forecasts upward because of expected AI-related demand growth. American Electric Power said this week it expects 63 gigawatts of incremental electricity load by 2030, driven largely by data centers. For crypto miners that survived the industry’s downturns, AI infrastructure is increasingly being viewed as a more predictable and potentially more profitable business than mining bitcoin alone. If you're reading this, you’re already ahead. Stay there with our newsletter .
