TeraWulf has undergone a dramatic revenue transformation in the second quarter of 2026, with high-performance computing (HPC) leases for artificial intelligence workloads now accounting for approximately 71% of total sales—up from 62% in the prior quarter The Block. The shift marks a decisive pivot for the company, which has traditionally been categorized as a Bitcoin miner.
The company's Bitcoin mining revenue collapsed by 73% during the same period, creating a stark contrast between its legacy operations and its emerging AI infrastructure business. Total quarterly revenue reached $44.8 million, with HPC leasing contributing the substantial majority The Block.
The 52% jump in HPC leasing revenue has widened TeraWulf's lead over its remaining cryptocurrency mining operations, which now represent less than 30% of the company's income stream. This revenue crossover positions TeraWulf among a small but growing cohort of former crypto mining firms that have successfully repurposed their energy infrastructure and data center expertise for the AI compute market.
However, the transition comes with notable timing constraints. While HPC leases now dominate reported revenue, contracted revenue from Anthropic—the AI company backed by Amazon and Google—is scheduled to commence only as new capacity is delivered from late 2027 onward CryptoSlate. This creates a temporal gap between current reported revenue composition and the full realization of contracted AI partnerships.
The 71% revenue share figure represents realized income rather than future contracted commitments, indicating that existing HPC customers—separate from the Anthropic arrangement—are already generating substantial cash flows. The distinction between current lease revenue and future contracted revenue suggests TeraWulf has built a diversified base of AI infrastructure clients beyond its headline partnership.
TeraWulf's infrastructure strategy leverages its existing relationships with power providers and its operational experience in energy-intensive computing. The company's facilities, originally designed for Bitcoin mining's continuous power demands, require modification for the different cooling, networking, and reliability specifications of AI training and inference workloads.
The revenue shift reflects broader market dynamics affecting the crypto mining sector. Declining Bitcoin mining economics—driven by network difficulty adjustments, halving events, and energy cost pressures—have compressed margins for pure-play miners. Simultaneously, demand for AI compute capacity has surged, creating pricing power for infrastructure providers with available power and operational expertise.
The 73% decline in mining revenue suggests TeraWulf has deliberately deprioritized cryptocurrency operations rather than merely experiencing market-driven fluctuations. The company appears to be reallocating capital and operational focus toward HPC infrastructure buildout, accepting near-term mining revenue reductions in exchange for longer-duration, potentially more stable AI lease contracts.
For investors and sector analysts, TeraWulf's quarterly results provide a case study in infrastructure repurposing. The company's transition demonstrates that specialized data center capabilities—particularly power procurement and facility management—can translate across compute-intensive industries, though successful execution requires significant capital deployment and technical adaptation.
The late 2027 timeline for Anthropic-specific revenue indicates that TeraWulf's current HPC revenue derives from other customers, suggesting a pipeline of AI infrastructure demand that extends beyond its highest-profile contracted relationship. This revenue diversification reduces dependence on any single counterparty while the company scales capacity to fulfill larger commitments.
TeraWulf's transformation raises questions for the broader crypto mining sector regarding optimal asset configuration. Firms with favorable power positions and geographic flexibility face similar strategic choices: maintain Bitcoin mining exposure, pursue hybrid models, or execute full transitions to AI infrastructure provision. The 71% revenue threshold crossed by TeraWulf suggests that for some operators, the pivot has already become the core business rather than a diversification strategy.