Meta announces a $14 billion Texas AI data center campus with BlackRock co-investment, expanding global infrastructure footprint in a low-cost power region.
Meta Platforms and BlackRock announced a partnership to develop a $14 billion data center campus in Texas focused on AI infrastructure. This represents Meta's largest external infrastructure collaboration to date and is intended to expand AI compute capacity to support the training and deployment of Meta's AI models across its products and services.
The project is structured to allow Meta to scale compute infrastructure while sharing the funding burden with institutional capital. Meta is contributing land and construction in progress valued at approximately $2.3 billion and retaining a 20% stake, which allows the company to maintain operational control and initial occupancy. BlackRock and its partner funds are funding most of the capital and taking 80% ownership. The planned 1 gigawatt of compute capacity is designed to support training and deployment of Meta's AI models across Facebook, Instagram, WhatsApp and other services, directly supporting the company's focus on AI-driven engagement and monetization. The campus is located in El Paso and is expected to reach planned capacity by 2028.
The partnership supports Meta's strategy of building large-scale AI infrastructure to support personalized experiences, advertising tools, and new products while using external capital to reduce the direct burden on the company's balance sheet. By holding a minority stake and providing management services while BlackRock funds most of the build cost, Meta gains access to substantial compute capacity with lower direct capital outlay than a fully owned project would require.
However, the venture presents several considerations for investors. The El Paso project relies on roughly $12.5 billion of debt raised by BlackRock funds, tying a core Meta AI initiative to a partner capital structure that could face refinancing or interest rate pressures over time. Concentrating 1 gigawatt of AI compute at a single campus introduces project execution, regulatory, and power supply risks, particularly as Meta competes with Alphabet and Microsoft for AI infrastructure resources. The long development runway and scale of spending raise concerns about margins and free cash flow if AI monetization occurs more slowly than expected. The use of third-party ownership structures for multi-gigawatt compute campuses, and any associated distributions, may not be fully reflected in high-level discussions of Meta's AI investments and could influence how investors evaluate future data center funding models.