TeraWulf (NASDAQ: WULF) is accelerating its transition toward large-scale artificial intelligence and high-performance computing infrastructure, with its latest quarterly results highlighting both expanding commercial commitments and the financial demands of building data center capacity. The company exceeded revenue expectations in Q3 2026, supported by HPC leasing activity, while major customer agreements and power commitments strengthened its longer-term development pipeline.

HPC Leasing Becomes the Primary Revenue Driver

TeraWulf generated $76.2 million in third-quarter revenue, surpassing the $61.8 million estimate included in the earnings report. HPC lease revenue contributed $63 million, accounting for approximately 83% of total revenue. This concentration demonstrates how significantly the company’s business model is shifting toward providing infrastructure for high-performance computing rather than relying exclusively on traditional digital asset mining activities.

The company reported 102 MW of operational critical IT capacity and another 336 MW under construction. Its Lake Mariner facility accounted for the reported operational and construction capacity, positioning the site as an important component of its expanding infrastructure platform.

However, adjusted EBITDA was reported at a loss of $12.8 million, compared with an estimated positive $12.8 million. This represents a significant difference from the revenue performance and indicates that higher leasing activity has not yet translated into the adjusted operating profitability anticipated by the market. Diluted EPS was negative $0.03, better than the estimated loss of $0.19.

Anthropic Agreement and Power Commitments Strengthen the Pipeline

One of the most significant developments was TeraWulf’s approximately 401 MW Anthropic lease, representing an estimated $19 billion in contracted revenue. The agreement highlights the growing demand for dedicated computing infrastructure from companies developing and operating advanced AI systems.

The company also reported 1,000 MW of contracted power at its Muskie project. Power availability is increasingly important in the data center industry because AI workloads require substantial electricity to operate large GPU clusters and associated cooling systems. Securing power commitments can therefore provide an important foundation for future capacity development, although contracted power should not be confused with operational IT capacity.

At Lake Mariner, 102 MW was already online, with another 336 MW under construction. TeraWulf affirmed its fiscal 2026 guidance for incremental critical IT capacity of approximately 250 MW to 500 MW annually. Achieving this target will depend on construction progress, equipment deployment, power infrastructure and customer delivery schedules.

The scale of these commitments illustrates the company’s ambition to participate in the expanding AI infrastructure market. Nevertheless, long-term contracted revenue is not equivalent to revenue already recognized, and the timing of future financial contributions will depend on the completion and commissioning of the underlying facilities.

Asset Monetization and Financing Support Expansion

TeraWulf ended the quarter with approximately $2.4 billion in cash, supported by several financing and asset-management developments. Among them was the sale of a 50.1% interest in Abernathy, which generated approximately $250 million in cash proceeds and a reported $95 million gain.

The company also secured a $400 million revolving credit agreement, including $150 million in accordion commitments. This arrangement provides additional financial flexibility as TeraWulf continues investing in data center development and infrastructure expansion.

Another notable earnings item was a $728.4 million fair-value gain associated with warrant and derivative mark-to-market adjustments. This accounting gain contributed to the reported net income of $709.9 million. However, fair-value adjustments can fluctuate with market conditions and do not necessarily represent cash generated from ordinary business operations. Consequently, the reported net income should be assessed separately from adjusted EBITDA and operating cash generation.

These developments demonstrate how TeraWulf is using asset monetization, financing arrangements and long-term customer contracts to support its transition. The company’s financial position provides significant reported liquidity, but the capital requirements associated with large-scale data center construction remain substantial.

Looking ahead, TeraWulf’s progress will depend on converting its contracted pipeline into operational capacity and recurring HPC lease revenue. Investors will be monitoring the delivery of the Anthropic agreement, the development of the Muskie power infrastructure, Lake Mariner’s construction schedule and the company’s ability to improve adjusted EBITDA. The central question is whether its expanding AI infrastructure portfolio can translate into sustainable operating profitability while maintaining sufficient financial flexibility to fund future growth.