TeraWulf (WULF) Q1 2026 earnings review
The Crossover: HPC Overtakes Bitcoin as TeraWulf Reinvents Itself
TeraWulf has officially crossed the Rubicon. In Q1 2026, High-Performance Computing (HPC) lease revenue ($21.0M) surpassed Bitcoin mining revenue ($13.0M) for the first time. Total revenue remained flat YoY at $34.0M, but the underlying business quality has entirely changed. However, this transition is expensive. GAAP net loss reached a staggering $427.7M, heavily driven by $216.3M in non-cash warrant liability changes and $67.1M in interest expense. The company sits on a war chest of $3.1B in cash, but it is paired with a massive $5.8B debt load. TeraWulf is no longer a crypto play; it is a highly-leveraged, execution-heavy AI infrastructure developer.
🐂 Bull Case
HPC lease revenue accelerated 117% sequentially. The 60 MW deployment for Core42 at Lake Mariner is energized, proving the company can execute complex infrastructure builds for top-tier tenants.
The portfolio has expanded to 2.3 GW across diverse regions. The recent acquisition of Hawesville, Kentucky brings 480 MW of immediate grid-connected power, solving the industry's biggest bottleneck.
🐻 Bear Case
Net debt sits at $2.7B. Interest expense was $67.1M in Q1 alone. TeraWulf has $1.48B in remaining adjusted cash CapEx just for the WULF Compute segment, leaving little room for execution missteps.
The transition involves painful accounting. A $25.7M impairment charge on old equipment and massive non-cash warrant liabilities keep GAAP earnings deeply negative, masking underlying operational cash flows.
⚖️ Verdict: ⚪
Neutral. Management has successfully executed a dramatic pivot to HPC, and the revenue mix proves it. But the sheer scale of the debt stack and the remaining capital required to build out the 2.3 GW pipeline require flawless execution over the next 24 months.
Key Themes
HPC Segment Margins Expanding
The HPC transition is paying off at the unit economics level. HPC segment margin expanded 152% QoQ, hitting $15.2M on $21.0M of revenue—an accelerating ~72% margin. Management is targeting a stabilized ~85% NOI margin on these contracts, showing that once the concrete is poured, the cash flow generation is highly lucrative.
Hawesville Acquisition Unlocks Immediate Growth
TeraWulf acquired the Justified Data campus in Hawesville, Kentucky. This is a massive win because it offers 480 MW of immediate grid-connected power in the MISO region. Management expects to sign a tenant here by Q2 2026, targeting a 2H 2027 delivery. Controlling ready-to-use power is TeraWulf's primary growth driver.
Macro Tailwinds: The Interconnection Bottleneck
Management explicitly highlighted that access to power has become the primary constraint in AI infrastructure. With utility interconnection queues severely backlogged, TeraWulf's strategy of acquiring brownfield sites with existing infrastructure (like former coal plants) allows them to bypass the queue and offer hyperscalers faster 'time-to-power'.
Bitcoin Mining is a Reversing Laggard
Bitcoin mining revenue plummeted 50% sequentially to $13.0M, down from $34.4M a year ago. The segment margin shrank to $5.0M. TeraWulf is aggressively cannibalizing this segment, taking impairment charges ($25.7M in Q1) to repurpose Lake Mariner infrastructure for HPC workloads. Investors should view mining as a terminal, runoff asset.
Mountain of Complex Liabilities
While management touts a '$3.1B liquidity position', the balance sheet is heavy. Total debt is $5.8B ($3.2B Senior Secured Notes, $2.5B Convertible Notes). Furthermore, warrant liabilities spiked to $1.06B, triggering a $216.3M non-cash fair value loss. This capital structure relies entirely on completing sites on time to start amortizing the massive interest burden.
Standardized Serial Delivery
TeraWulf is moving away from bespoke builds to a standardized design. The CB-3, CB-4, and CB-5 facilities use a repeatable module. This specific technological approach compresses execution risk, allowing the company to target sequential delivery of capacity (e.g., CB-3 in May, CB-4 in Q3, CB-5 in Q4).
Liquidity vs Implied CapEx Burn
Despite having $3.1B in cash and restricted cash, the company faces daunting CapEx. The WULF Compute segment alone requires $1.48B in remaining CapEx to finish CB-3, 4, and 5. The company's positive narrative about being fully funded contradicts the reality that any construction delay or cost overrun could force a return to the capital markets.
Other KPIs
Stable sequentially compared to $(4.7)M in Q1 2025, but representing a massive reversal from the deeply negative $(23.1)M printed in Q4 2025. The normalization of SG&A expenses and the ramp of high-margin HPC revenue were the primary factors stopping the bleeding.
Decelerating from an adjusted basis, but still incredibly elevated compared to $46.6M in Q1 2025. This includes $101.4M in stock-based compensation, meaning actual cash burn on corporate overhead is much lower, but shareholder dilution remains a real cost of retaining talent.
Guidance
Stable. The company reaffirmed its strategy of targeting up to half a gigawatt of new contracted capacity per year. At $8M-$10M per MW build cost, this implies ongoing annual CapEx requirements of $2.0B to $5.0B.
Accelerating timeline execution. The 42 MW facility is nearing completion, keeping the company on schedule to meet its commitments to its second major tenant, Fluidstack. CB-4 and CB-5 remain on track for late 2026.
Management expects to announce a tenant for the newly acquired 480 MW Hawesville site in the coming quarter. Achieving this will be a major catalyst to prove demand extends beyond their legacy New York site.
Key Questions
Margin Normalization Timeline
HPC segment margins expanded to ~72% this quarter. How long until the segment reaches the targeted 85% NOI margin, and what fixed costs need to be absorbed to bridge that gap?
Debt Servicing Sustainability
With $67M in interest expense this quarter and remaining CapEx of $1.5B+ across active projects, at what MW utilization threshold does the business become truly free cash flow positive after debt service?
Mining End-of-Life Plan
Given the $25.7M impairment charge and rapid revenue decay in the Bitcoin mining segment, what is the exact timeline for completely decommissioning mining operations at Lake Mariner to reclaim that power capacity for HPC?