- Meta Compute looks to oversee massive AI computing infrastructure expansion
- Initiative reports directly to Mark Zuckerberg and operates at the top corporate level
- Tens of gigawatts are planned this decade and hundreds are expected in the future
Meta has established a new internal organization to oversee the expansion of its computing infrastructure for advanced AI tools.
The new Meta Compute initiative operates at the top level of the company and reports directly to CEO Mark Zuckerberg, who says it plans to deploy tens of gigawatts this decade.
Over a longer time frame, the company expects capacity to scale to hundreds of gigawatts, far exceeding traditional data center growth patterns.
The timing of Meta Compute is notable, as the company will spend about $72 billion on artificial intelligence by 2025, but the financial payoff remains unclear.
Meta has emphasized that these investments aim to deliver economic benefits in the areas where data centers are being built.
This issue has become more sensitive as communities question the impact of large plants on electricity prices and water consumption.
The new organization brings together software, hardware, network and facilities planning under one umbrella.
Meta has indicated that this structure is intended to ensure that hardware and software decisions remain aligned, which is necessary as AI workloads place different demands on systems compared to previous cloud services.
Meta Compute will be jointly led by Santosh Janardhan and Daniel Gross, with responsibilities split between execution and long-range planning.
Janardhan continues to oversee deeply technical areas including system architecture, internal silicon development, software layer and the global data center fleet.
Gross will focus on defining future computing requirements, building supply chains capable of delivering multi-gigawatt-scale hardware, and developing planning models that account for industry shifts and resource constraints.
Together, their mandate reflects an attempt to treat power, land, equipment and networks as a single coordinated problem.
“Today we’re establishing a new top-level initiative called Meta Compute,” Zuckerberg wrote in a post about Thread.
“Meta plans to build tens of gigawatts this decade, and hundreds of gigawatts or more over time. How we engineer, invest and collaborate to build this infrastructure will be a strategic advantage.”
At the same time, Meta Compute separates long-term capacity strategy from day-to-day data center operations, which continue under existing infrastructure teams.
This split suggests that Meta is trying to avoid reactive expansion driven only by short-term demand.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews and opinions in your feeds. Be sure to click the Follow button!
And of course you can too follow TechRadar on TikTok for news, reviews, video unboxings, and get regular updates from us on WhatsApp also.
