Meta Launches Dedicated 'Meta Compute' Division to Manage Massive AI Data Center Expansion

Meta Launches Dedicated 'Meta Compute' Division to Manage Massive AI Data Center Expansion

January 12, 2026

In a major strategic move underscoring the immense computational demands of the artificial intelligence race, Meta Platforms Inc. has established a new internal division, "Meta Compute," to oversee its ambitious plans to build multiple AI data centers at a gigawatt-plus scale. This reorganization signals a new phase of infrastructure investment as tech giants compete to secure the power and hardware necessary to train and deploy next-generation AI models.

The new division will be co-led by Santosh Janardhan, Meta's head of global infrastructure, and Daniel Gross, a prominent AI investor and former CEO of Safe Superintelligence (SSI). Gross joined Meta after the company's attempt to acquire SSI, founded by ex-OpenAI chief scientist Ilya Sutskever, was unsuccessful. As part of the hiring package, Meta also acquired Gross's investment fund, NFDG.

Meta's commitment to this infrastructure build-out is staggering in both scale and financial terms. CEO Mark Zuckerberg stated, "Meta is planning to build tens of gigawatts this decade, and hundreds of gigawatts or more over time." This expansion is backed by a series of massive financial commitments. The company recently formed a $27 billion joint venture with Blue Owl Capital to fund its 2-gigawatt "Hyperion" data center campus in Louisiana, with Blue Owl covering 80 percent of the cost. Furthermore, Meta has warned investors of "notably larger" capital expenditures for 2026, following a record 2025.

To meet its power requirements, Meta has proactively secured agreements for up to 6.6 gigawatts of nuclear energy through partnerships with companies like TerraPower, Oklo, and Vistra. The company has also embarked on a spending spree for external cloud capacity, reportedly signing deals exceeding $10 billion with Google Cloud, $14.2 billion with CoreWeave, $3 billion with Nebius, and is in talks for a potential $20 billion contract with Oracle.

The creation of Meta Compute formalizes the company's bifurcated approach to AI dominance. While a separate superintelligence lab, led by Scale AI's Alexandr Wang—whose recruitment involved Meta acquiring a 49 percent stake in Scale AI for approximately $14.8 billion—focuses on AI software development, the new division will manage the physical and strategic underpinnings. Janardhan will continue overseeing technical architecture and data center operations, while Gross will lead long-term capacity planning and supplier partnerships. They will be supported by Dina Powell McCormick, a former Goldman Sachs executive and government advisor, who joined this week to manage partnerships with governments and sovereign entities for infrastructure financing and deployment.

This move highlights a critical trend in the technology sector: AI advancement is increasingly constrained by physical infrastructure—namely, data center space, energy supply, and specialized hardware. Meta's centralized, well-funded push into gigawatt-scale computing sets a new benchmark for the industry and intensifies competition for limited global resources, from power grids to semiconductor supply chains. The success of Meta Compute will be a decisive factor in the company's ability to keep pace in the escalating global AI contest.

Source: datacenterdynamics