Crusoe, specializing in cloud computing and infrastructure for artificial intelligence, raised $3.9 billion in a new investment round at a company valuation of $30.9 billion. The round was led by Atreides Management, Mubadala Capital, and Valor Equity Partners. Other participants include Nvidia, Founders Fund, GIC, Qatar Investment Authority, Radical Ventures, and TPG. For Crusoe, this represents a significant increase in valuation: in October 2025, the company received $1.38 billion at a valuation of $10 billion. Thus, in less than a year, the figure more than tripled. TechCrunch notes that with the new valuation, Crusoe has entered the ranks of the highest-valued private companies operating in the AI infrastructure sector.
One of the areas where the company plans to direct the raised funds is the modular data centers Spark. They are manufactured at Crusoe's own facilities, after which the ready blocks are transported by freight and installed in locations with significant volumes of available electricity. This approach allows for reducing the deployment time of computing infrastructure and minimizing the amount of construction work directly on site. Spark is planned to be used primarily for inference – processing user requests with already trained AI models. Such tasks usually require fewer computing resources than during the training of large models, which opens up the possibility of placing the corresponding infrastructure in a more compact format.
At the same time, Crusoe is not abandoning the construction of traditional large data centers. The company is already collaborating with OpenAI, Microsoft, and Oracle, and in early September, it signed a five-year agreement with financial company Jane Street for the supply of computing power worth approximately $13 billion. Simultaneously, Crusoe is considering further business expansion: in August 2026, Axios reported that the company was discussing a potential initial public offering with Goldman Sachs and Morgan Stanley. Thus, Crusoe is developing several infrastructure formats at once – from large data processing centers to modular sites designed to place AI services closer to sources of available electricity.
Related articles
