The Project
The El Paso data center will be Meta’s largest AI training facility to date. Key details include:
- Total investment: $14 billion jointly funded by Meta and BlackRock
- Site size: 1,000 acres in El Paso, Texas
- Power capacity: 1.5 gigawatts, including dedicated renewable energy
- Compute capacity: Designed for next-generation AI model training
- Jobs: 3,000 construction jobs, 800 permanent positions
- Timeline: Construction to begin Q1 2027, first phase operational by 2028
The site was selected for its proximity to existing fiber optic infrastructure, favorable energy costs, and the absence of zoning complications that have delayed similar projects in other states.
Why Texas
Texas has become the epicenter of AI infrastructure development. The state offers several advantages:
- Deregulated energy market: Competitive electricity prices and the ability to negotiate direct power purchase agreements
- Favorable business climate: No state income tax and streamlined permitting processes
- Fiber connectivity: Major internet exchange points in Dallas and Houston provide low-latency connectivity
- Land availability: Large parcels of land suitable for data center campuses
The El Paso project joins a growing list of major AI infrastructure investments in Texas, including Oracle’s $20 billion campus in Austin and Tesla’s AI training center in Austin.
Industry Context
The Meta-BlackRock partnership reflects the growing role of financial institutions in AI infrastructure. Traditional data center financing involved tech companies building and operating their own facilities. The BlackRock partnership represents a new model where financial institutions provide capital and expertise in exchange for long-term returns.
BlackRock’s involvement signals that AI infrastructure is becoming an asset class of its own. The firm manages over $10 trillion in assets and has been increasing its allocation to digital infrastructure. The Meta partnership represents its largest single investment in AI.
For Meta, the partnership reduces the capital burden of building AI infrastructure while ensuring access to the compute needed for its AI research. The company is developing Llama 4, its next-generation open-source language model, and needs massive compute capacity for training.