Meta (META.US) signs contract with Crusoe to purchase 1.6GW of data center computing power, further increasing investment in AI infrastructure.

According to sources, Meta Platforms (META.US) has reached a new cooperation agreement with data center developer Crusoe to obtain large-scale artificial intelligence (AI) computing resources from multiple data centers built by the latter to support the company's continued expansion of its AI strategy. Sources said that Meta has signed contracts to purchase computing capacity from two data centers located in Childress, Texas and Warrenton, Missouri, operated by Crusoe. As the negotiations are not public yet, the sources requested to remain anonymous. It is reported that Meta will acquire a total computing capacity of around 1.6 gigawatts (GW) from the two projects. The specific amount Meta will pay for this and when the computing resources will be delivered were not disclosed by the sources. Both Meta and Crusoe spokespersons declined to comment on this. In recent years, as the AI competition intensifies, Meta, along with tech giants such as Microsoft Corporation (MSFT.US), Alphabet Inc. Class C (GOOGL.US), and Amazon.com, Inc. (AMZN.US), is accelerating its focus on securing data centers, power, and computing resources to meet the demand for large model training and inference. Currently, Meta's largest AI infrastructure project is a nearly 4,000-acre data center campus in Louisiana, expected to provide up to 5 gigawatts of computing capacity in the future. The company has also pledged to invest at least $60 billion in the AI infrastructure sector over the next few years to solidify its competitive position in the field of artificial intelligence. Established in 2018, Crusoe specializes in AI data center construction and operation and has rapidly grown to become an important player in the US AI infrastructure sector in recent years. In addition to Meta, the company has established partnerships with tech giants such as Oracle Corporation (ORCL.US), Microsoft Corporation, and Alphabet Inc. Class C, to build large AI data center campuses. However, Crusoe has faced challenges in some of its project expansions. Earlier reports indicated that due to failed negotiations with Alphabet Inc. Class C, Crusoe's data center project in Wyoming faced setbacks, and Oracle Corporation had also considered sourcing computing resources from the project. Earlier this month, Crusoe disclosed that the company's contracted data center capacity has reached 4.9 gigawatts, with an overall project pipeline exceeding 40 gigawatts. This includes both projects with signed lease agreements and data centers in advanced development stages or under negotiation with potential clients. Among the signed projects are the Abilene data center campus in Texas, serving Oracle Corporation and OpenAI, as well as projects related to Microsoft Corporation. Furthermore, local media previously reported that a third data center project near Amarillo, Texas has been leased by Alphabet Inc. Class C. Last week, Crusoe also revealed that two additional data center projects in Texas and Missouri have secured tenants, but did not disclose the specific client names. Meta's acquisition of approximately 1.6 gigawatts of additional computing resources once again reflects the escalating competition among tech giants for AI infrastructure. With the continued expansion of AI model sizes, rapid growth in inference demand, and increasingly strained power resources, data center capacity has become one of the most critical strategic resources in the tech industry. At the same time, Meta is facing increasing pressure to demonstrate to investors that its significant investments in AI can generate tangible returns. As the parent company of Facebook, Instagram, and WhatsApp, Meta recently started offering a subscription service for its AI chatbot Siasun Robot & Automation, seen by the market as an important step in exploring the commercialization of AI. Analysts believe that as the company continues to expand its computing resources and drive AI product monetization, the ability to form a sustainable profit model in the AI business in the coming years will be a focus of market attention.