TeraWulf’s Q1 loss hits $427M as HPC leasing becomes majority of revenue
HPC lease revenue rose 117% QoQ to $21M, while Bitcoin mining revenue fell 50% to about $13M.
TeraWulf’s Q1 2026 results put hard numbers on its miner-to-AI pivot: HPC leasing is now the majority of revenue. The trade-off is a sharply wider net loss, with limited detail in the packet on what drove the $427 million figure.
Key Takeaways
- TeraWulf recorded a $427 million net loss in Q1 2026, widening from a $61.4 million loss a year earlier.
- Total Q1 revenue was $34 million, with HPC lease revenue at $21 million, roughly 60% of sales after a 117% quarter-on-quarter increase.
- Bitcoin mining revenue declined 50% to around $13 million during the quarter.
- The HPC line was tied to 60 MW of operational critical IT capacity at Lake Mariner leased to Core42.
TeraWulf’s Q1 Print: AI/HPC Becomes the Majority as Losses Widen
TeraWulf’s Q1 2026 print shows the pivot is no longer theoretical. Revenue totaled $34 million, and the HPC leasing line contributed $21 million, making it the majority of reported sales.
The problem for traders is that the income statement did not stabilize alongside that mix shift. The company posted a $427 million net loss for the quarter versus a $61.4 million loss in Q1 2025. With no itemization in the provided materials for what drove the loss, the headline number becomes the near-term overhang even as the AI/HPC line scales.
In equity terms, the tape reflected that tension. WULF shares closed down 2.6% on the day, despite being described as up more than 105% year-to-date and up over 30% over the past month.
Inside the Revenue Mix: $21M From Core42 at Lake Mariner
The revenue mix is the cleanest signal in the quarter. HPC lease revenue jumped 117% quarter-on-quarter to $21 million, while Bitcoin mining revenue fell 50% to around $13 million. That is the business flip in one line: leasing is now the larger contributor, and mining is the smaller one.
The disclosed driver also makes concentration risk easy to frame. TeraWulf attributed the quarter’s HPC revenue to 60 megawatts of operational critical IT capacity at its Lake Mariner site leased to Core42. In practice, that means incremental megawatts delivered and leased are the key datapoint for whether the leasing line can keep compounding fast enough to offset mining volatility and decline.
Long-Dated Leasing Pipeline: Fluidstack/Google and 2026–Q4 2026 Delivery Targets
Management is leaning on long-dated contracted revenues to support the AI infrastructure narrative. TeraWulf previously announced a 25-year lease deal with Fluidstack, backed by Google, worth around $9.5 billion in contracted revenues, described as an expansion of an earlier 10-year commitment.
The timeline matters as much as the headline contract value. The company said it is coordinating infrastructure delivery with Fluidstack and Google, with additional capacity buildings on track for delivery in 2026. It also said the Abernathy joint venture, a 168 MW HPC project under a 25-year lease, remains on track for delivery in the fourth quarter of 2026.
CFO Patrick Fleury framed the financing approach around matching duration, saying, “Our capital structure is designed to align long-term financing with contracted cash flows, supporting disciplined growth while maintaining financial flexibility.”
Signals Traders Can Track Next: Cash, Build Deliveries, and Mining vs. Lease Trajectory
The next catalyst is not another narrative update. It is basic disclosure and execution.
First, traders need itemization of what drove the $427 million net loss in the next filing or earnings materials, including whether impairments, non-cash charges, or financing items dominated the quarter.
Second, the real operational scoreboard is delivery. Updates on additional capacity buildings slated for 2026, tied to Fluidstack/Google coordination, will indicate whether the current 60 MW base can expand quickly enough to keep HPC leasing outgrowing the mining decline.
Third, the Abernathy joint venture’s Q4 2026 target is a dated milestone. Any progress markers or slippage signals will likely matter more than broad “AI pivot” language.
Finally, TeraWulf said it ended the quarter with approximately $3.1 billion in cash, but the packet provides no balance sheet breakdown. Clarification and corroboration around that figure, including restrictions and intended uses, is a key risk-control datapoint for how the buildout gets funded.
The Pivot Is Showing Up in Revenue—But the Income Statement Is Still the Risk
I treat this quarter as confirmation that TeraWulf’s mix has already flipped, with HPC leasing now doing the heavy lifting on revenue while mining shrinks. The threshold that matters is whether new MW deliveries turn that $21 million line into a repeatable ramp rather than a single-customer step function.
The real test is whether the next set of disclosures explains the $427 million loss in a way the market can normalize. If that loss is largely non-cash or one-time, the setup starts to look structural rather than narrative-driven. If it is recurring economics, the AI pivot stays tradable, but the multiple will keep fighting the income statement.
