TeraWulf is transitioning from bitcoin mining to contracted HPC leasing, with 60 MW already online at Lake Mariner and more capacity slated for 2026. The opportunity is real, but so are the losses, leverage, and valuation risk.
TeraWulf (WULF) is no longer a clean read as a bitcoin miner and not yet a mature data center landlord. The investment case sits in that awkward but potentially lucrative middle ground. In Q1 2026, the company reported $34.0M of revenue, with $21.0M from HPC leasing and $13.0M from digital asset revenue, marking the first quarter where HPC leasing was meaningfully reflected in results. That mix shift matters because management is trying to replace volatile mining revenue with long-duration, credit-backed infrastructure revenue.
The bull case rests on three hard facts. First, 60 MW of critical IT HPC capacity at Lake Mariner was energized and generating revenue as of March 31, 2026. Second, management said CB-4 and CB-5 remain on schedule for delivery in Q3 and Q4 2026, while the Abernathy joint venture targets Q4 2026 delivery. Third, the company had about $3.1B of cash and restricted cash at March 31, 2026, giving it unusual liquidity for a company still in buildout mode.
The bear case is just as real. Q1 2026 GAAP net loss was $427.6M, loss per share was $(1.01), adjusted EBITDA was still negative at $(4.1)M, and total debt stood at $5.20B at fiscal 2025 year-end. Annual revenue in 2025 was only $168.5M, yet the company carried an EV/Revenue multiple of 78.4x. That is venture-style valuation sitting on top of project-finance leverage. When execution is smooth, markets forgive that. When execution slips, the math gets ugly fast.
For a balanced, moderate-risk investor with a medium-term horizon, WULF looks like a speculative Buy rather than a core holding. The company has real assets, real contracts, and visible HPC ramp. It also has a thin margin for error because valuation already prices in substantial success. The stock works best for investors who want exposure to the AI-power bottleneck theme and can tolerate high volatility, a 4.261 beta, and a business model still moving from formation to delivery.
The company’s current identity is best understood through its transition. Historically, TeraWulf built and operated bitcoin mining infrastructure. By Q1 2026, that legacy base was still producing revenue, but the business mix had shifted sharply. HPC lease revenue of $21.022M represented more than 60% of Q1 revenue, while digital asset revenue fell to $12.990M from $34.405M in Q1 2025. In plain English, the company is trying to turn power access and site control into contracted AI infrastructure cash flow.
Management framed the shift clearly on the Q1 2026 earnings call. Paul Prager said, “We are fundamentally a power company that builds digital infrastructure, not the other way around.” That line is more than branding. It explains why TeraWulf keeps emphasizing sites, interconnection, generation, and utility relationships rather than just mining rigs or generic colocation capacity.
The asset base now centers on Lake Mariner, the Hawesville, Kentucky site, the Morgantown acquisition in Maryland subject to regulatory approval, and the Abernathy joint venture in Texas. Lake Mariner is the operating proof point. Kentucky is the near-term expansion story with immediate power availability. Morgantown is the strategic option on a constrained Mid-Atlantic market. Abernathy adds another contracted buildout path.
TeraWulf currently operates through two economically distinct activities: digital asset mining and HPC leasing. The difference between them is not cosmetic. Mining is commodity-like and exposed to bitcoin price, network difficulty, and power economics. HPC leasing is infrastructure-like and tied to contract duration, tenant credit, and delivery milestones.
In Q1 2026, HPC leasing generated $21.022M of revenue versus none in Q1 2025. Management said this was the initial ramp of long-term customer agreements at Lake Mariner. The lease with Core42 commenced in March after CB-2 achieved Ready for Service, and the company said all 60 critical MW of Core42 capacity had been delivered by quarter-end. That is the most important operating fact in the report because it proves the model can move from slide deck to invoice.
Digital asset revenue in Q1 2026 was $12.990M, down from $34.405M in Q1 2025. That decline shows both the volatility of the legacy business and management’s deliberate repurposing of mining footprint toward higher-value HPC workloads. On the call, Paul Prager said, “Mining served its purpose. It enabled us to build infrastructure, monetize power and develop operational expertise. But the future of this platform is contracted long-duration compute infrastructure.”
Segment economics also point in the same direction. CFO Patrick Fleury said the as-reported quarterly HPC segment profit margin was about 50%, but adjusting for tenant fit-out revenue, pre-revenue operating costs at WULF Compute, and development costs across uncontracted sites, the margin would have been about 85% in Q1. Even if investors haircut that adjusted figure, the message is clear: mature HPC leasing should be structurally more attractive than mining.
The risk is that the old business is shrinking before the new business is fully scaled. That creates a messy transition period where reported revenue can look flat, GAAP losses can look ugly, and valuation can look detached from current earnings power. WULF is living in that exact phase now.
Get AI research on any stock
Instant reports, daily intelligence, and an AI analyst in your pocket.
TeraWulf’s flagship product is not a machine or a software stack. It is contracted, power-backed HPC capacity at Lake Mariner. That campus is the company’s operating showcase and the place where the strategy is being tested in public.
As of March 31, 2026, Lake Mariner had 60 MW of critical IT HPC capacity energized and generating revenue. Management said the second data hall in CB-2 came online during the quarter, bringing all Core42 capacity into service. In effect, Lake Mariner has become the first real proof that TeraWulf can convert power access into recurring lease revenue rather than just mining output.
The next steps at the site are equally important. Management said CB-3 remained on schedule, with defined scope expected to be completed by the end of May, while CB-4 and CB-5 remained on track for delivery in the third and fourth quarters of 2026. Those milestones matter because the market is not paying for the current 60 MW alone. It is paying for the belief that Lake Mariner can keep stacking contracted capacity on time.
That comment from Paul Prager gets to the heart of the product. TeraWulf is not selling generic square footage. It is selling powered, engineered, tenant-ready capacity in a market where power availability is the choke point. That makes the product less like traditional real estate and more like a grid-linked industrial asset with digital tenants attached.
The quality of the flagship product also depends on contract structure. The 10-K states that the HPC leases are long-term data center lease agreements, with fixed payments recognized on a straight-line basis over the noncancellable term and variable payments for power delivery billed at actual utility rates. That structure supports revenue visibility while limiting direct commodity exposure on the power pass-through component.
TeraWulf’s competitive edge is built around power control, site selection, and integrated infrastructure delivery. That sounds simple, but in the current AI data center race, simple is where the moat lives. Plenty of companies can market AI exposure. Fewer can secure large-scale power, permit the site, finance the build, and deliver tenant-ready capacity.
Management’s language has been consistent on this point. Paul Prager said, “We are not pursuing power as an input cost. We are structuring power as a core asset.” He also said, “Securing queue position alone is no longer enough,” arguing that procurement capability, delivery credibility, and financial assurance are becoming more important differentiators. That thesis fits the broader market reality: the bottleneck has shifted from demand to deliverability.
The company’s vertical integration is another advantage. TeraWulf owns and operates infrastructure rather than acting as a thin intermediary. That gives it more control over timing, design changes, and tenant coordination. Nazar Khan highlighted that campuses at this scale require alignment across power infrastructure, equipment procurement, construction delivery, and customer hardware deployment. In a market where hardware roadmaps keep changing, that coordination matters.
There is also a capital advantage, at least relative to smaller peers. As of March 31, 2026, the company had about $3.1B of cash and restricted cash. Fleury also said the parent held about $300M of available unrestricted cash at quarter-end, rising to about $1.5B after April equity raises. That does not erase leverage risk, but it does improve the odds that TeraWulf can keep building while weaker operators stall.
The catch is that competitive advantage in this market expires quickly if execution slips. A power moat is real, but only if the company can energize capacity on schedule and keep counterparties happy. In infrastructure, the moat is less a castle wall and more a bridge. It works only while it keeps carrying traffic.
Operations are the center of the WULF story. The company is building large-scale campuses where power infrastructure, customer fit-out, and hardware deployment have to line up with almost military precision. That is why management keeps returning to execution rather than broad market hype.
At Lake Mariner, the operating update was concrete. Core42’s 60 MW was fully delivered by the end of Q1 2026. CB-3 was on track for scope completion by the end of May, and CB-4 and CB-5 were on track for Q3 and Q4 2026 delivery. Those are the milestones that matter most over the next few quarters because each building brought online shifts revenue mix further toward contracted HPC.
The company is also expanding beyond New York. Since year-end, management said it added the Hawesville, Kentucky site, described as a large-scale campus with immediate power availability and significant expansion potential. The company is targeting 480 MW online in the second half of 2027 for Kentucky. It also repaid the $100M draw on the bridge credit facility after quarter-end and terminated that facility.
Morgantown, Maryland is the more strategic but less certain operating lever. Management said the acquisition remained subject to regulatory approval, with a FERC decision expected in mid-summer. Prager described the site as attractive because of its location in the Washington, D.C. and Northern Virginia corridor and because it is a brownfield industrial site rather than a greenfield farm conversion. That is a meaningful siting advantage if approved.
Abernathy adds another layer. Management said the joint venture is being delivered under a lump sum EPC agreement with Hypertec, with Fluidstack leading development and construction. As of March 31, 2026, the JV had about $1.4B of gross cash, or $1.0B net of reserve and lockbox accounts, with $0.4B of CapEx spent and $0.9B remaining. That is a large project, and it underlines how much of WULF’s future depends on construction discipline.
WULF sits at the intersection of two markets moving in opposite directions on quality. Bitcoin mining remains volatile and cyclical. AI and HPC infrastructure is attracting premium valuations because demand for powered compute capacity is strong and supply is constrained by grid access. TeraWulf is trying to migrate from the first market into the second without losing its footing.
The company’s own results show why that migration matters. Q1 2026 revenue was essentially flat YoY at $34.0M versus $34.4M, but the composition changed dramatically. HPC lease revenue went from $0 to $21.0M, while digital asset revenue dropped to $13.0M from $34.4M. The revenue line barely moved, yet the quality of that revenue improved. Markets tend to pay more for contracted infrastructure cash flow than for mining output tied to bitcoin and network difficulty.
Management said it entered 2026 with 522 critical IT MW of contracted HPC capacity and a gross 2.9-GW multi-regional platform. It also reaffirmed a strategy targeting 250 MW to 500 MW of new contracted capacity annually. Those figures frame the addressable opportunity inside the company’s own pipeline. The market is effectively valuing WULF on how much of that contracted capacity turns into operating revenue on schedule.
Industry positioning also matters. TeraWulf’s peer set still includes bitcoin miners such as Marathon Digital (MARA), Riot Platforms (RIOT), CleanSpark (CLSK), Hut 8 (HUT), Bitfarms (BITF), Cipher Mining (CIFR), and Bit Digital (BTBT). But strategically, WULF is moving closer to hybrid infrastructure names that monetize power and data center assets through AI and HPC hosting. That peer migration is one reason the stock can trade at a valuation that looks absurd through a miner lens and still attract buyers.
Like what you're reading?
Get full access to AI-powered research reports, market analysis, and portfolio tools.
TeraWulf’s customer base is shifting from mining-pool counterparties and bitcoin-linked economics toward enterprise and AI compute tenants. The named customer evidence in the record includes Core42, Fluidstack, and Google-linked deployment activity at Lake Mariner. That customer profile is important because it changes both revenue quality and operating demands.
Core42 is already revenue-producing. The company said the lease with Core42 commenced in March 2026 after CB-2 reached Ready for Service, and all 60 MW of Core42 capacity had been delivered by quarter-end. That makes Core42 the clearest proof point that TeraWulf can serve large, sophisticated counterparties rather than just run self-mining infrastructure.
Management also described active coordination with Fluidstack and Google on final energization and hardware deployment. Nazar Khan said contract terms in newer deals have moved from 10 years to 15 years, which is a notable signal. Longer contract duration can improve revenue visibility and financing support, but it also raises the bar on design flexibility because hardware changes over a 15-year term are almost guaranteed.
The customer profile creates both strength and risk. On the positive side, these are larger, more creditworthy counterparties than a typical mining customer base. On the negative side, concentration risk rises because a few large tenants can dominate near-term HPC revenue. For now, WULF is not selling to thousands of small users. It is landing a handful of very large elephants, and elephants do not turn quickly.
The competitive landscape around WULF is best split into two camps. The first camp is public bitcoin miners competing for power, hardware, and capital. The second camp is emerging AI and HPC infrastructure operators competing for tenants, utility relationships, and shovel-ready sites.
Against pure miners, WULF’s transition is a strategic advantage. Legacy mining economics have become tougher after the halving, and revenue tied to bitcoin production is inherently less predictable than lease revenue. WULF’s Q1 2026 mix, with HPC leasing contributing more than 60% of revenue, already looks different from miners still dominated by hashprice exposure.
Against infrastructure-oriented peers, the picture is more mixed. WULF has a credible power-and-site story, but it is still proving out execution at scale. The company’s market cap was about $10.50B against trailing annual revenue of $168.1M, and EV/Revenue was 78.4x. That valuation implies investors already view WULF as more than a miner. The burden now is to prove it deserves to be valued alongside premium infrastructure stories rather than speculative transitions.
Institutional sponsorship is a plus here. Institutional ownership stood at 82.404%, with 15 of 20 tracked institutions increasing positions. Vanguard held 33.6M shares and BlackRock held 28.1M shares. That level of sponsorship does not guarantee success, but it does show that sophisticated capital has bought into the transition narrative.
Short interest adds another layer. Short interest was 25.82% of float with a short ratio of 4.03. That is high enough to matter. It reflects skepticism around valuation and execution, but it also creates fuel for sharp upside if the company keeps hitting build milestones. WULF is the kind of stock where the operating story and the trading story can feed each other, for better or worse.
The macro backdrop for WULF is dominated by one fact management repeated several times: power is the bottleneck. On the Q1 2026 call, Prager said, “The constraint is not GPUs. It is power.” That is the right lens. AI demand matters, but the monetizable edge sits with companies that can secure, permit, and energize large-scale capacity in constrained regions.
That backdrop favors operators with existing sites and utility relationships. TeraWulf’s Hawesville site has immediate power availability, Morgantown sits in a constrained Mid-Atlantic corridor, and Lake Mariner is already operational. In a market where interconnection queues and transmission limits slow new supply, brownfield or power-advantaged sites become strategic assets.
There is also a financing macro at work. TeraWulf’s buildout is capital intensive, and the company’s ability to raise equity and project capital is central to the story. Management said approximately $1.2B of equity had been raised year-to-date to help fund the Kentucky project. That is supportive, but it also means WULF remains tied to capital market openness. If risk appetite tightens, funding costs and dilution pressure become more painful.
Regulatory and permitting factors matter as well. Morgantown requires FERC approval, and Cayuga is in the planning permit process. These are not side issues. For a company monetizing power access, regulatory timing can be as important as customer demand. The geopolitical angle is less about cross-border trade and more about domestic grid policy, energy infrastructure, and the strategic value of U.S.-based AI capacity.
About $3.1B of cash and restricted cash at March 31, 2026 gives TeraWulf unusual liquidity, but $5.20B of debt and a 4.261 beta leave little room for execution mistakes.
Unlock the full analysis
Premium members get the complete breakdown — pick rationale, financial metrics, and recent earnings detail.
Get Full Access →Q1 2026 revenue reached $34.0M, yet a $427.6M GAAP net loss and $(4.1)M adjusted EBITDA show the business is still far from consistent profitability.
Unlock the full analysis
Premium members get the complete breakdown — pick rationale, financial metrics, and recent earnings detail.
Get Full Access →Management says CB-4 and CB-5 remain on schedule for Q3 and Q4 2026 delivery, while the Abernathy JV targets Q4 2026, setting up a meaningful HPC ramp.
Unlock the full analysis
Premium members get the complete breakdown — pick rationale, financial metrics, and recent earnings detail.
Get Full Access →At 78.4x EV/Revenue on 2025 sales of $168.5M, TeraWulf trades like a venture-stage infrastructure story rather than a mature data center operator.
Unlock the full analysis
Premium members get the complete breakdown — pick rationale, financial metrics, and recent earnings detail.
Get Full Access →The report’s fair value sits at $35, with the stock looking attractive only if Lake Mariner and the next buildouts keep converting power access into contracted lease revenue.
Unlock the full analysis
Premium members get the complete breakdown — pick rationale, financial metrics, and recent earnings detail.
Get Full Access →TeraWulf (WULF) is one of the more interesting transition stories in the public market. The company has already proven that its HPC strategy can generate real revenue, with $21.0M of lease revenue in Q1 2026 and 60 MW energized at Lake Mariner. It also has a visible build pipeline, meaningful liquidity, and a management team that is talking less about dreams and more about megawatts, leases, and delivery.
But this is still a speculative infrastructure build, not a finished compounder. Net losses remain large, leverage is high, and valuation assumes the company keeps converting power assets into contracted revenue with few major stumbles. That is why the stock earns a Buy, not a Strong Buy, for a moderate-risk investor.
The medium-term setup is attractive because the business mix is improving and the market theme is powerful. The discipline is to remember what kind of stock this is. WULF is not a cheap asset play. It is a premium execution story. When premium execution stories execute, they can run hard. When they miss, the market stops being patient in a hurry.
Get AI-powered research reports, daily market intelligence, and a personal analyst in your pocket.
Get Full Access →Not ready to subscribe? ·
TeraWulf is being priced with bitcoin-miner baggage even after landing a 20-year Anthropic lease tied to roughly $19 billion of contracted revenue. The story has shifted toward AI infrastructure ownership, and the market still looks early to that re-rating.
TeraWulf Inc. (WULF) jumps after announcing a 20-year lease with Anthropic that could generate about $19 billion in contracted revenue. The deal, plus a majority stake sale in its Abernathy JV, strengthens the company’s AI infrastructure pivot and gives investors a clearer long-term growth story.
D-Wave Quantum is showing real booking momentum, but revenue, profitability, and valuation still lag far behind the hype. The stock looks like a Hold as the company works to convert early commercial wins into durable cash flow.