TeraWulf: High-Density HPC Contracts Drive Long-Term Infrastructure Value

Summary TeraWulf’s latest $9.5 billion, 25-year HPC deal with Fluidstack anchors predictable, multi-year revenue streams, and the joint venture structure spreads capital costs and limits WULF’s upfront ￰0￱ retains 51% control, managing expansion pace and operational ￰1￱ two Fluidstack deals provide up to $954 million in annual revenue potential, with projected NOI margins of 70% and a conservative asset value of $1.83 ￰2￱ client concentration is a risk, WULF's scalable, next-gen facilities could attract hyperscalers, positioning WULF for sustainable upside and further contract ￰3￱ (NASDAQ: WULF ) has struck gold with its WULF Compute ￰4￱ the first Fluidstack deal was announced back in August, WULF has seen over 200% upside, and last week’s extension of the deal through a new $9.5 billion , 25-year joint venture at TeraWulf’s Abernathy, Texas site, just put WULF on another multi-year and multibillion-dollar revenue ￰5￱ new agreement adds 168 MW of critical IT load, built to HPC-grade ￰6￱ main takeaway from my last WULF coverage , which came just after the first Fluidstack deal was extended by an additional 160 MW of capacity back in August, when WULF was already up 180% on the news at the time, was that WULF still had room for a run.

I upgraded WULF to a Buy in that ￰7￱ has seen an additional 72% upside ￰8￱ has done a few things right to become one of the market favorites in the HPC race among Bitcoin ( BTC-USD ) ￰9￱ of them is prioritising long-life power and cooling infrastructure rather than owning a large, depreciating GPU fleet. I believe WULF is getting these back-to-back deals because of the technical buildout of its ￰10￱ through the engineering choices at the Lake Mariner buildout, which includes liquid-cooled Tier III-grade design and N+1 cooling redundancy, the site was designed from the ground up for high-density compute with closed-loop liquid cooling, dual 345 kV transmission lines, redundant low-latency fiber, and direct grid access to ￰11￱ setup gives WULF lower energy costs, high uptime, and scalable capacity that can handle AI and HPC workloads without expensive ￰12￱ engineering choices make WULF’s facilities plug-and-play ready for clients like Fluidstack who need stable, low-latency, high-efficiency ￰13￱ website The site at Abernathy is the primary location for the latest $9.5 billion Fluidstack deal, which will be adding 168 MW of critical IT load under a joint venture with Fluidstack, built to HPC-grade standards with liquid cooling, scalable transmission, and redundant fiber ￰14￱ also owns 51% of the venture and holds exclusive rights for future expansion at the same location, meaning the joint venture can double capacity under similar ￰15￱ WULF is executing aggressively, as an investor, a key question for the longer term is whether WULF’s setup is positioning it for a multi-year run or could this be an overextension?

I believe the setup of TeraWulf's HPC deals structure protects the company from overextension because the new joint venture with Fluidstack spreads capital costs and limits how much WULF must commit ￰16￱ of the HPC spending goes into power and infrastructure, not GPUs, which means less exposure to rapid depreciation and better asset retention. TeraWulf's sites were built for modular expansion, so capacity can scale only when utilization supports ￰17￱ the contracts are customer-funded, with Fluidstack locking in long-term commitments that give predictable cash flows before full ￰18￱ a 51% controlling interest, WULF keeps operational control and can manage expansion pace ￰19￱ this is a setup that stays asset-heavy where value holds and light where risk ￰20￱ is a setup that contributes to long-term net operating income (NOI) visibility and capital efficiency.

Long-Term Financial Impact of TeraWulf’s HPC Deals The way the two Fluidstack deals are structured results in predictable NOI for ￰21￱ the announced first Fluidstack deal, TeraWulf projects an 85% NOI ￰22￱ the second deal, the expected NOI margin is 70%. Combined, the two Fluidstack deals include 168 MW from the latest $9.5 billion Abernathy project, and 200 MW from the initial Lake Mariner colocation deal plus the 160 MW added at Lake Mariner, which gives ~530 MW in current HPC ￰23￱ the combined ~530 MW across both Fluidstack deals, the numbers show how the HPC business could reshape TeraWulf’s value ￰24￱ benchmarks for high-density AI hosting put annual revenue between $1.5 million and $2 million per ￰25￱ is reported that AI data centers in North America currently rent out space to colocation clients at an average price of $150 per kW monthly recurring charge, which would translate to about $1.8 million per MW a ￰26￱ the mid-range figure of $1.8 million per MW gives roughly $954 million in annual revenue potential for WULF’s ~530 MW HPC ￰27￱ also has the potential to attract even larger hyperscaling clients for colocation partnerships who could pay up to $200 to $300 per kW a month, especially since the tech infrastructure at TeraWulf’s sites is being built to be next-gen and plug-and-play for these hyperscalers, with liquid cooling systems and high-density rack ￰28￱ Lake Hawkeye Data Campus is also being developed exclusively for HPC and AI data center hosting, with exclusive rights to develop up to 400 MW of HPC capacity, and 138 MW of power expected to be ready for service in Q2 of ￰29￱ adds more revenue potential and platform scale to WULF ￰30￱ NOI margins projected to be around 70%, that translates to about $667.8 million in potential NOI (derived from applying a 70% margin to total projected revenue of $954 million for the 530 MW).

WULF’s 51% share comes to roughly $340.6 million in attributable NOI once both sites reach steady-state ￰31￱ that the 51% share applies only to the second deal, which WULF is developing as a joint venture with ￰32￱ the first deal, WULF has potential for a higher site-level NOI margin (the company projects 85%) and likely a much higher NOI share since it is a colocation deal rather than a joint ￰33￱ to stress test this valuation conservatively, I'm treating both deals under a 51% ownership scenario and 70% ￰34￱ stay even more conservative, assume WULF achieves only a 60% NOI margin conversion and 50% utilization at the ￰35￱ cuts attributable NOI to about $146 million (calculated as $954 million x 0.6 x 0.5 x 0.51).

Applying an 8% capitalization rate, which factors in the early-stage infrastructure risk, gives an implied asset value of about $1.83 billion for the HPC business alone (derived by dividing NOI by the cap rate). For an industry context, Fitch Ratings assigned a 9% cap rate to Equinix ( EQIX ), a leading data center peer worth noting in this comparison. I applied an 8% cap rate for WULF, one percentage point lower than EQIX’s 9%; WULF will likely be operating with stronger operational certainty and less churn risks because of its longer-term HPC contracts and perhaps superior asset quality (based on the cooling and rack design the sites technical specifications show), which collectively reduce risk and justify a tighter ￰36￱ with the very conservative setup like above, where we assume TeraWulf achieves only a 60% NOI margin conversion and 50% utilization at both HPC sites, it still shows that even under modest execution, WULF’s HPC business retains meaningful asset value and a solid intrinsic valuation base for the HPC ￰37￱ Only risk I can envision for now is that TeraWulf has announced Fluidstack as the tenant in its two biggest HPC deals.

I think it would be a better strategy to ink more contracts with diverse arrays of clients and possibly some hyperscalers. Today, IREN ( IREN ) inked a multi-year cloud contract deal with Microsoft, showing how large hyperscalers are entering this ￰38￱ in TeraWulf's deals, Google backstopped $3.2 billion of the initial Fluidstack lease obligations and an additional $1.3 billion for the new joint venture, which inherently adds some credit quality, there would still be a more balanced risk profile for WULF in terms of revenue visibility and counterparty concentration if its HPC portfolio were diversified among several blue-chip ￰39￱ TeraWulf’s HPC business is on the verge of becoming a multi-year growth ￰40￱ back-to-back Fluidstack deals put WULF on a billion-dollar revenue ￰41￱ company’s focus on long-life power and cooling infrastructure, and customer-funded expansions reduces capital risk while maximizing operational ￰42￱ main risk is client concentration, but continued contract diversification could mitigate ￰43￱ investors, the takeaway is clear: WULF is potentially positioned for sustainable upside as the technical infrastructure and scalable HPC footprint attract hyperscalers and premium colocation contracts.