Meta Platforms has intensified its pursuit of artificial intelligence infrastructure by finalising new agreements with data centre developer Crusoe to secure approximately 1.6 gigawatts of computing capacity across two facilities in the United States. The move, confirmed through a Bloomberg report citing individuals familiar with the matter, positions the social media giant to expand the computational resources essential for training and deploying next-generation AI models at scale.
The capacity Meta will acquire spans Crusoe data centres located in Childress, Texas, and Warrenton, Missouri. While financial terms and the precise timeline for operational deployment remain undisclosed, the combined 1.6-gigawatt envelope equals the electricity required to power roughly 1.2 million American households simultaneously. This agreement represents one of the most significant recent infusions of AI compute capacity secured by a major technology company.
Gigawatts Have Become the New Currency in AI Competition
The technology industry’s artificial intelligence race has fundamentally shifted from a model-centric contest to an infrastructure battle measured in gigawatts. Success now depends on who can secure sufficient electricity, data centre space, and computing hardware to sustain increasingly complex AI workloads. One gigawatt alone can supply power to approximately 750,000 U.S. homes, making Meta’s 1.6-gigawatt commitment a substantial strategic asset.
Meta CEO Mark Zuckerberg has positioned AI as the central pillar of the company’s future trajectory, with heavy investments directed toward AI assistants, autonomous agents, recommendation systems, and large language models. The company recently announced a $600 billion commitment to U.S. infrastructure and jobs over three years, underscoring the scale of capital required to support its AI ambitions.
Crusoe Emerges as Critical Player in AI Infrastructure Ecosystem
Crusoe has rapidly evolved into a pivotal force within the AI infrastructure landscape. Earlier this month, the company reported contracting nearly 4.9 gigawatts of AI infrastructure capacity across its data centre projects and cloud platform, with a development pipeline exceeding 40 gigawatts. Founded in 2018 by Chase Lochmiller and Cully Cavness, Crusoe pivoted from cryptocurrency mining to AI infrastructure as demand for computing power surged.
The startup’s distinctive energy-first approach enables it to control multiple layers of the infrastructure stack—from energy sourcing and facility construction to computing deployment and cloud operations. This integrated model has accelerated Crusoe’s buildout pace at a time when AI compute demand consistently outpaces available supply. The company now serves Microsoft, Oracle, OpenAI, and numerous AI enterprises through its Crusoe Cloud platform.
Meta’s Broader Infrastructure Strategy Signals Long-Term AI Commitment
The Crusoe agreement complements Meta’s expanding portfolio of infrastructure partnerships. The company recently expanded its AI cloud alliance with CoreWeave to a $21 billion deal spanning 2027 through 2032, secured long-term access to NVIDIA’s advanced AI chips, and finalised its first AI-enabled data centre agreement in India with Reliance Industries for a 168-megawatt facility in Jamnagar, Gujarat.
These coordinated investments reflect a strategic reality: building competitive AI systems requires unsustainable levels of infrastructure if companies rely solely on internal capacity. Long-term partnerships with data centre developers enable hyperscalers to access specialised energy infrastructure, geographic diversity, and specialised cooling systems that would be prohibitively expensive to develop independently.
As AI models become more computationally intensive, the industry is witnessing a fundamental transformation in how technology giants approach infrastructure. The ability to secure gigawatt-scale capacity has become as critical as model architecture itself, reshaping competitive dynamics across the AI ecosystem.
