Summary TeraWulf is pivoting from crypto mining to AI/HPC data center infrastructure, securing major deals with Core42 and Google/Fluidstack. TeraWulf's partnerships could generate up to $15.7B in revenue over a decade, with Google backstopping $3.2B in lease obligations and taking a 14% equity stake. Aggressive buildout targets and innovative deal structures position TeraWulf for rapid growth, with analysts forecasting a near 100% CAGR in revenue over two years. Despite risks like share dilution and potential demand drops, Google's involvement and strong revenue prospects justify a Buy rating on TeraWulf. Investment Thesis Just like many of its crypto miner peers, TeraWulf Inc. ( WULF ) also started off this year with one goal in mind—to diversify its data center infrastructure footprint away from crypto mining towards the highly lucrative AI/HPC (high-performance) workload space. The Maryland-based crypto miner struck a deal with Core42, the UAE-based cloud service provider, to deliver brand-new data center infrastructure equipment to Core42 by the end of the year, resulting in a billion dollars of potential revenues over a decade. As if that was not enough, the crypto miner also struck a deal with Alphabet Inc. ( GOOG ) (GOOGL) and Fluidstack that sees the search giant become a major investor in TeraWulf’s transformation into an AI data center infrastructure company. The nature of these deals has enough potential to alter the course of TeraWulf’s history and possibly win newer clients down the road, which makes TeraWulf look very attractive at current levels. I am recommending a Buy rating on TeraWulf, as I explain in the thesis below. What Makes TeraWulf Stand Out In The Rapid Buildout I want to start my analysis of TeraWulf’s prospects with an admission. This recent slate of new deals that markets have witnessed over the past few months in the AI DC (data center) space is extremely intriguing. The structure of these deals has been innovative in the sense that it interlocks partners with common synergies, allowing every stakeholder to progress towards achieving the deal’s eventual objective—delivery of hundreds of megawatts of data center capacity fitted with accelerators. It doesn’t matter if they’re merchant GPUs like NVIDIA Corporation ( NVDA ) or custom accelerators like Google’s TPUs. These deals definitely carry some risk due to the size and scale of the deals, which I will explain in a later section. But aside from that, the inventive deal composition definitely does a better job at aligning all stakeholders towards a common goal rather than the typical debt financing that we witnessed during the 90s dot-com bust. That being said, TeraWulf stands out in these deals to turn crypto miners' current infrastructure footprints into powerful DCs that can handle HPC workloads. Let me start by briefly summarizing TeraWulf’s foray into the hypergrowth market for supporting HPC workloads with mission-critical DC infrastructure. So far, a majority of TeraWulf’s $144M worth of TTM revenues comes from mining cryptocurrencies out of its Lake Mariner facility in upstate New York which has a capacity of ~245 MW based on its quarterly filings . The asset utilization rate, when it comes to the revenues generated per MW of infrastructure used, seemed low when the opportunity of serving the AI market presented itself to TeraWulf and its peers. When TeraWulf announced the deal with Core42 last year, they mentioned that they were targeting $1.5M per megawatt hour on a previous quarterly call with investors . That was a significant improvement to the future state of its asset utilization as compared to its crypto mining business. If TeraWulf has hit its milestones from the Core42 deal, it should currently be in the process of delivering 70 MW of turnkey DC infrastructure. So far, TeraWulf has not updated investors on the delivery of the turnkey DC infrastructure to Core42 yet, but that could be announced any time soon. Then, through the last two months, Alphabet’s Google and Neocloud Fluidstack expanded their partnership with TeraWulf, raising the stakes. The expanded partnership accomplished two particular objectives, the second more important than the first. Firstly, TeraWulf reserved the potential to see $6.7B in contracted revenue, with the option of benefiting by another $9.3B if Google/Fluidstack exercised lease extension options. In return, TeraWulf would have to provide ~360 MW of DC capacity at its Lake Mariner campus, which it does not have today. That means it would have to build this from the ground up. But the delivery milestones that TeraWulf had to hit seemed aggressive, with ~40 MW expected to come online in H1 CY26, while the rest of the ~320 MW was expected to come online by H2 CY26. This requires a scale of resources, both technical and financial, that is massive to justify the rapid build-out of the 360 MW of DC capacity for Google/Fluidstack. This can only be supported by meaningfully ramping up financial leverage. Here’s how TeraWulf’s CFO, Patrick Fleury, framed the structure of the deal and reasons for getting financial leverage to build out the capacity: The cost for these data centers is really driven by 2 things. It's driven by design for your specific customer and/or hardware, and then it's driven by schedule. The faster and bigger you are, the more it’s going to cost, because you just have to throw more people [resources] at it. Additionally, he added: Then you go forward to Fluidstack & Google, 450 [megawatts] gross. About 360-ish net [megawatts]. $6.7B of contract value. Core42 was roughly a billion in contract value. Also, a 10-year deal with additional options with Fluidstack and Google could add another $9B of revenue. $15.7B of potential revenue. Google, alongside that deal, is backstopping about $3.2 billion of the lease obligations, which is roughly again, roughly 50% of the $6.7B total over the first 10 years. They are also, in return, taking a 14% stake in the equity. Exhibit A: TeraWulf’s deal with Google and Fluidstack (Company presentation) Unlike the dot-com era debt financing, where vanilla debt was raised by vendors to build out dot-com infrastructure, the recent deals between Google/Fluidstack and TeraWulf involve a complex structure that involves Google backstopping TeraWulf’s funding that it needs to source to finance the 360 MW build-out. Google is also doubling its stake in TeraWulf, now holding ~14% in its indirect vendor, TeraWulf. (I say indirect vendor because Google has a relationship with TeraWulf via Fluidstack) . The “Google Put” in this relationship effectively means Google has cemented its interest in the future prospects of TeraWulf by not only backstopping $3.2B in future lease obligations but also holding 14% of the crypto mining company. Bloomberg said that Terawulf has reportedly already raised $3B in debt to finance their commitments to Google and Fluidstack, which is a positive note in the progress that TeraWulf is making. If TeraWulf pulls off this scale of DC buildout, they would unlock billions in potential revenue to the tune of a minimum of $6.7B spread out over 10 years, which is significant to the future prospects of TeraWulf. Investors should also note that these kinds of partnerships with Core42 and Google/Fluidstack pave the way for TeraWulf to acquire newer clients due to the high-profile nature of the deals. Exhibit B: TeraWulf’s commitments for next year and beyond. (Company presentation) TeraWulf Has Room To Surge Higher I am going to look over current forward-looking revenue multiples of 29x forward revenue (EV/revenue forward) because that is factoring in FY25 revenues of $206M, growing 47% y/y. Analysts are penciling in a near 100% CAGR in revenue growth rates for TeraWulf over the next two years, as seen below. Exhibit C: Revenue estimates for TeraWulf for the next two years. (YCharts) Yet when I compare TeraWulf’s enterprise value of ~$6B versus its FY27 revenues, it indicates a forward revenue multiple of just 7.3x, which is incredibly cheap for TeraWulf. Compare TeraWulf to Iren, another peer that trades at 7.4x FY27 revenues of $2.2B, growing at a CAGR of 41%. Risks & Other Factors To Note Investors may be concerned with share dilution rates, which currently are tracking at a pace of +30% CAGR over the past 2 years, but I believe TeraWulf may have reached peak share dilution rates because of the additional financing TeraWulf is getting from the Google backstop. More importantly, TeraWulf’s shares can be badly hit if its clients suddenly stop spending on TeraWulf’s GPU infrastructure. This is possible if there is a recession or if the end customer demand for AI services starts to drop, pushing customers like Google to rein in spending, impacting TeraWulf. This does pose a meaningful risk to investors, and I would advise investors to invest in TeraWulf based on their risk appetite. Takeaway Google’s involvement in TeraWulf at the investor level and at the financing level is a significant deal for the crypto miner, allowing it to get investment-grade debt, which it can use to build DC capacity for its clients. The revenue opportunity for TeraWulf due to these deals is also significant. I am recommending a Buy rating on TeraWulf.
