The relentless pursuit of artificial intelligence has sparked an unprecedented demand for computational power, driving a surge in infrastructure spending across the tech sector. While industry giants unveil their ambitious AI strategies, a surprising cadre of companies, once primarily focused on cryptocurrency mining, has quietly ascended to become pivotal players in this high-stakes arena. These firms leverage their established energy infrastructure to deliver the critical, power-intensive data center capacity essential for the AI revolution.
Companies like TeraWulf, Applied Digital, Iren, Core Scientific, and Cipher Digital are redefining their operations. Their most valuable asset isn’t just a knack for complex computing; it’s privileged access to utility power contracts. These agreements allow them to rapidly energize the enormous electrical needs of advanced AI computing facilities – a foundational requirement for success in this burgeoning sector. The robust power connections, initially secured for their energy-hungry crypto mining operations, now attract top-tier customers in desperate need of high-performance computing space.
This strategic pivot represents a remarkable transformation. From a market grappling with the volatility of digital currencies, these entities have seamlessly transitioned into the boundless opportunities presented by the AI boom. Nick Giles, a senior research analyst at B. Riley Securities, who closely tracks several of these publicly traded firms, highlights the unforeseen scale of their success. His team’s analysis reveals an astonishing collective market capitalization leap for 11 leading former crypto-mining firms, rocketing from approximately $2.1 billion in late 2022 to an astounding $48.5 billion today. This, Giles asserts, represents a “winning scenario” for early investors.
Brian Dobson, a managing director and head of technology equity research at Clear Street, echoes this sentiment, calling it a “mindblowing jump in valuation.” However, this meteoric rise on the stock market now demands tangible results. These companies must execute complex, multi-billion-dollar projects to construct state-of-the-art AI facilities for the world’s most demanding tech firms. This new endeavor is significantly more expensive and intricate than their previous mining operations, hinging on successful integration with power grids already straining under the immense demand from data centers. The cumulative energy consumption of this sector in the coming years is projected to rival that of several major American cities, underscoring the critical role of robust energy infrastructure.
The stakes are incredibly high. Failure to meet construction deadlines could trigger substantial financial penalties, potentially undermining the economics of these massive projects. For investors monitoring the energy market, understanding the power demand and grid reliability is paramount.
Forging Ahead as AI Energy Infrastructure Providers
The transition from cryptocurrency mining to AI data centers is not without its technical challenges. While crypto mining facilities operate at hundreds of megawatts, AI data centers demand an even higher standard of operational resilience. These facilities require minuscule tolerances for outages and necessitate significant investment in sophisticated equipment, including robust backup generators and extensive cooling systems to manage the intense waste heat generated by intensive AI computations. This shift elevates the complexity and capital expenditure for these developers.
Dobson emphasizes the industry’s evolving focus: “The theme of 2025 was securing contracts with major customers. 2026 is about expanding those agreements and, critically, executing on those deals.” Investors are keenly watching for successful project delivery.
A significant unknown remains the potential impact of construction delays or operational issues. Such setbacks could trigger financial penalties or even allow customers to withdraw from project commitments. Given the colossal costs associated with AI data center development, often running into billions of dollars for large campuses, any misstep could be financially devastating. While most of these former crypto-mining companies are publicly traded, the level of transparency regarding the intricate details of their customer contracts can vary.
Dobson candidly acknowledges the reality of large-scale construction: “There will be delays; there always is in construction.” An illustrative example surfaced late last year when CoreWeave, itself a former crypto miner that has evolved into a $40 billion AI cloud provider, disclosed during an earnings call that it was “affected by temporary delays related to a third-party data center developer, who is behind schedule” on construction. CEO Michael Intrator attributed the broader data center building boom to “overwhelming the supply chains.” This news led to a 10% drop in CoreWeave’s stock at the time. Equity analysts at Northland Securities later identified Core Scientific as the likely delayed partner, though Intrator confirmed in a subsequent February earnings call that these facility delays had since been resolved.
High-Stakes Ventures with Varied Transparency
Despite the inherent risks, the industry has largely seen lucrative deals overshadowing any public setbacks. Applied Digital and Cipher Digital currently lead among the former crypto miners in accumulating significant data center portfolios, boasting 5 gigawatts and 4.1 gigawatts, respectively, in active and planned capacity. To contextualize this scale for energy market participants, New York City’s summer power demand is approximately 10 gigawatts, highlighting the immense energy footprint these companies are developing.
Both firms have hinted at substantial upcoming agreements. Cipher Digital recently unveiled a 15-year commitment with an undisclosed “investment-grade hyperscale tenant,” while Applied Digital reported being in “advanced discussion on three sites and 900 megawatts.” However, detailed disclosures regarding the potential ramifications of execution stumbles from these companies have been limited.
TeraWulf stands out for its transparency. Last summer, following its announcement of a deal for Fluidstack and Google to expand to 360 megawatts at its Lake Mariner campus in Western New York, the Maryland-based company filed public documents with the Securities and Exchange Commission. These filings revealed a critical clause: if the project faces delays of 180 days or more, Fluidstack retains the right to terminate the lease, potentially voiding Google’s $3.2 billion rent guarantee. “180 days sounds like a comfortable amount of time, but it’s not that long when we’re talking about mega-construction,” notes B. Riley analyst Giles, who nevertheless expresses confidence in TeraWulf’s delivery capabilities. The company’s shares, which traded below a dollar in 2023, now hover around $15, making it one of Giles’ top investment picks.
The lending market, meanwhile, has signaled increasing confidence in the trajectory of these former crypto miners, extending debt at progressively more favorable rates. Cipher Digital, for instance, secured $1.4 billion at a 7.1% interest rate in November last year for a Texas data center project with Fluidstack and Google. In a testament to improving market sentiment, the company obtained a $2 billion loan at an even lower 6.1% rate in February this year to finance another Texas facility for Amazon. Similarly, Applied Digital borrowed $2.15 billion at 6.8% in March for an Oracle-dedicated data center. This marks a notable improvement from the 9.3% rate it received for a $2.35 billion loan in November last year for a facility being developed for CoreWeave, which carries a lower credit rating than Oracle.
Furthermore, there’s a growing industry consensus that even if these former crypto-mining firms encounter delays, their major clientele might be more forgiving due to the overarching scarcity of AI computing power nationwide. Paul Golding, a senior digital infrastructure analyst at Macquarie, observes, “While in some instances there have been delays on delivering some infrastructure, the customers have not changed any of the terms.” He attributes this leniency to the market dynamic: “The scarcity of power and infrastructure has created a dynamic where to go somewhere else would take far longer potentially than just waiting for the incremental catch up.” For investors, this implies a degree of resilience in the face of operational hurdles, driven by an insatiable demand for AI-ready energy infrastructure.
