en.Wedoany.com Reported - July 2 news, US AI data center startup Crusoe is in talks to raise approximately $3 billion to support the expansion of AI computing infrastructure. The financing is still in progress, with the final valuation yet to be determined. Investors estimate the company could be valued at around $30 billion, a significant increase from the approximately $10 billion valuation in October last year.
Crusoe's business growth is directly tied to AI computing supply contracts. The company has established computing supply relationships with major tech firms such as Meta Platforms and Oracle, catering to demand for GPU clusters, AI cloud services, and data center infrastructure. Large model training and inference require long-term, stable computing pools, which in turn demand GPU servers, high-speed networks, power distribution systems, racks, cooling equipment, data center operations, and cloud platform scheduling capabilities. To continue expanding delivery scale, Crusoe must align server procurement, facility construction, power access, network architecture, and customer lease contracts. Once the $3 billion financing is completed, it will provide stronger capital support for such heavy-asset construction and further enhance its project acquisition capabilities in the AI infrastructure market.
The company initially focused on oilfield associated natural gas power generation and cryptocurrency mining before gradually transitioning to AI data centers and cloud computing services. The construction of US AI data centers is currently impacted by grid connection cycles, substation construction, land approvals, cooling resources, and server delivery timelines. Companies that can quickly organize power, facilities, and computing equipment are more likely to secure large client contracts. Crusoe has disclosed nearly 5 gigawatts of contracted computing power capacity and maintains a project pipeline exceeding 40 gigawatts. Such capacity reserves determine its ability to continuously deliver large-scale GPU computing power to major model companies and cloud service providers.
Meta's computing procurement contract with Crusoe involves two data center projects in Childress, Texas, and Warrenton, Missouri, with a combined capacity of approximately 1.6 gigawatts. For clients, they are purchasing long-term, callable AI computing services; for data center operators, each major contract must correspond to a clear construction schedule, server delivery, power supply assurance, network debugging, and operations system. The investment intensity of AI computing data centers far exceeds that of standard facilities, with GPU servers having high unit costs and power density, imposing higher demands on switching networks, cooling systems, and power redundancy. Crusoe's current financing discussions reflect that AI cloud vendors are competing for the next phase of computing orders through capital, energy resources, and engineering delivery capabilities.
If the financing is successfully completed, Crusoe will gain more funds for data center construction, GPU cluster deployment, and AI cloud service expansion. The company's valuation could rise from approximately $10 billion to around $30 billion, indicating that investors view power capacity, data center project pipelines, and long-term computing contracts as core assets of AI infrastructure companies. As clients like Meta and Oracle continue to expand their AI clusters, competition among data center companies will increasingly focus on power access speed, facility delivery efficiency, network architecture capabilities, and large-scale GPU operations.
