Wall Street is reassessing the role of Bitcoin mining companies. Bernstein, in its latest research report, states that these companies are no longer solely reliant on mining revenue, but are gradually becoming key suppliers in the expansion of AI data centers, leveraging readily available electricity, land, and data center resources.
17 agreements signed in two years
Bernstein stated that over the past two years, Bitcoin mining companies have signed 17 related agreements totaling over $110 billion and provided approximately 6 gigawatts of power resources to companies such as Google, Amazon, Microsoft, Nvidia, and CoreWeave.
The agency pointed out that this capacity accounts for about 10% of the total number of AI data centers under construction in the United States. With the demand for AI computing power continuing to rise, the ability to provide readily accessible power faster is becoming a key area of competition in the industry.
Covering two mining companies
Bernstein has begun covering TeraWulf and Cipher Digital, both of which have received an "Outperform" rating. The research report suggests they have significant potential to benefit from the transformation of AI infrastructure.
Bernstein predicts that TeraWulf's AI-related revenue could reach $1.7 billion by 2030. Its partnerships with Fluidstack and Google are seen as key drivers. During the same period, its EBITDA margin is projected to be close to 84%.
Cipher Digital's clients are primarily hyperscale cloud service providers. Bernstein projects the company's AI revenue will reach $1.2 billion by 2030, with an EBITDA margin approaching 93%.
Long-term contracts bring stable cash flow
The research report mentions that these companies prefer to adopt a data center hosting model, which involves leasing facilities with existing power supply to customers under long-term contracts and setting "take-or-pay" clauses to increase revenue certainty.
Bernstein stated that the project financing market currently covers 75% to 85% of the construction costs for such facilities, while financing rates are typically lower than the returns offered by the underlying contracts. This is a key reason why investors are paying attention to this model.
In a broader context, the competition for electricity in the AI industry is intensifying. As infrastructure bottlenecks beyond chips become increasingly prominent, companies that once secured electricity resources in advance for mining are now beginning to gain higher valuations on another front.
