Meta establishes ‘Meta Compute,’ plans multiple gigawatt-plus scale AI data centers : US Pioneer Global VC DIFCHQ SFO NYC Singapore – Riyadh Swiss Our Mind

Tens to hundreds of gigawatts of compute in the decades to come

Meta has set up a new division to manage its growing AI data center ambitions, dubbed Meta Compute.

“Meta is planning to build tens of gigawatts this decade, and hundreds of gigawatts or more over time,” CEO Mark Zuckerberg said.

Meta Mesa

– Meta

The effort will be co-led by Meta’s head of global infrastructure Santosh Janardhan and businessperson Daniel Gross.

Gross was the CEO of Safe Superintelligence (SSI), an AI business started by former OpenAI chief scientist Ilya Sutskever, who joined Meta after the social networking giant failed to acquire SSI. Meta also bought out Gross’ investment fund NFDG, as part of his hiring.

Zuckerberg has spent heavily on recruiting AI talent, including spending approximately $14.8 billion on a 49 percent stake in data labeling firm Scale AI and hiring its CEO, Alexandr Wang, to lead a new superintelligence lab.

While Wang is focused on the AI layer, the new division will focus on the infrastructure behind it. Janardhan will continue to manage Meta’s technical architecture, silicon, and data center operations, with Gross leading a new group dedicated to long-term capacity planning, supplier partnerships, and business modeling.

They will work with Dina Powell McCormick, a former Trump advisor and Goldman Sachs alum, who this week joined “to work on partnering with governments and sovereigns to build, deploy, invest in, and finance Meta’s infrastructure.”

Meta has already announced a significant data center spend, in October warning investors that capex costs will be “notably larger” in 2026 than the record 2025. Alongside its own gigawatt data center construction projects, Meta reportedly signed a $10bn+ cloud deal with Google in August, followed by a $14.2bn CoreWeave deal last September, a $3bn Nebius contract in November, and is in talks with Oracle for a $20bn contract.

To fund its 2GW Hyperion data center, the company entered into a $27bn joint venture with Blue Owl Capital, with the latter company covering 80 percent of the cost.

Earlier this month, Meta announced three new partnerships with nuclear companies to power its growing data center fleet: Two with small modular reactor developers, TerraPower and Oklo, and one with retail electricity giant Vistra. Combined, the deals could provide access to up to 6.6GW of nuclear power from a mix of advanced SMRs and existing large-scale reactors.

https://www.datacenterdynamics.com/en/news/meta-establishes-meta-compute-plans-multiple-gigawatt-plus-scale-ai-data-centers/