Meta and BlackRock partner for $14B gigawatt-scale AI campus in El Paso

This partnership formalizes the 'off-balance-sheet' financing model Meta is using to fuel its massive infrastructure buildout without traditional CAPEX bloat. By leveraging BlackRock’s capital, Meta can scale 1-gigawatt sites—essential for next-tier training and inference—while protecting its own cash flow. For the streaming and video ecosystem, this massive reserve of compute reinforces Meta’s ability to dominate AI-driven content recommendations and ad targeting or pivot into providing wholesale compute to third-party developers. Watch the 2028 go-live date as a benchmark for when Meta’s 'Meta Compute' marketplace could reach a critical mass of surplus supply to challenge tier-one cloud providers.

The El Paso venture follows Meta’s October 2025 milestone with Blue Owl Capital, which utilized a similar 80/20 ownership split to finance the 'Hyperion' data center in Richland Parish, Louisiana. Per Seeking Alpha, that project’s scope was recently expanded to 5 gigawatts with a $50 billion construction budget. Bloomberg News reported in July 2026 that the total investment for the Louisiana site could surpass $250 billion when accounting for the procurement of high-end AI chips. These massive commitments are part of Mark Zuckerberg’s 'AI-first' strategy, which has pushed Meta’s 2026 capital expenditure guidance to a range of $130 billion to $145 billion, according to company earnings data from July 2026. Simultaneous with its hardware buildout, Meta is maturing its revenue-generating AI products. On July 9, 2026, the company launched Muse Spark 1.1, a multimodal reasoning model designed for agentic tasks. As reported by CNBC, this marks Meta’s first foray into paid AI infrastructure, with developer API access priced at $1.25 per million input tokens. Per Fierce Network (July 2026), these infrastructure moves align with a broader 'Meta Compute' initiative. This internal unit aims to monetize data center surplus by selling raw compute capacity and hosted models, positioning Meta to compete directly with AWS, Google Cloud, and specialized GPU-rental neoclouds.