Meta Platforms is doubling down on its pursuit of artificial general intelligence (AGI), announcing plans to build several massive data centers over the next few years, with investments potentially reaching hundreds of billions of dollars. The company’s bet: supercharged infrastructure is the key to staying ahead in the global AI race.
Superintelligence labs and next-gen clusters
The new projects, dubbed Prometheus and Hyperion, are multi-gigawatt data centers designed to power Meta’s future AGI models. Prometheus is expected to be operational by 2026, while Hyperion is being built to scale up to an eventual capacity of 5 gigawatts.
The company also revealed plans to build multiple “titan clusters,” each reportedly the size of a large section of Manhattan. Meta aims to become the first AI research lab to bring a gigawatt-scale supercomputing cluster online, according to analysis by industry publication SemiAnalysis.
These efforts fall under Meta’s recently formed Superintelligence Labs, an internal division created to consolidate the company’s AI research following setbacks in its open-source Llama 4 model and the departure of several key team members.
Funding the future with ad revenues
Despite investor concerns about rising costs, CEO Mark Zuckerberg reassured stakeholders that Meta’s core ad business — which generated nearly $165 billion in revenue last year — remains strong enough to support the company’s ambitious infrastructure buildout.
The company raised its 2025 capital expenditure guidance to $64–72 billion, with much of the increase earmarked for AI expansion. This includes continued funding for Meta AI, image-to-video generation tools, and its smart glasses portfolio.
A high-stakes talent war
Zuckerberg’s vision isn’t just about silicon — it’s also about people. In recent weeks, Meta has aggressively recruited top AI engineers, including former Scale AI CEO Alexandr Wang and GitHub’s ex-chief Nat Friedman, both now leading Superintelligence Labs.
This talent grab comes on the heels of a $14.3 billion investment into Scale AI, and is viewed as part of Meta’s effort to keep pace with rivals like OpenAI and Google, both of which have ramped up hiring and product development in the AGI space.
While the road to superintelligence remains uncertain, Meta is positioning itself to shape the future of AI through relentless infrastructure growth, deep-pocketed investments, and a bold willingness to lead — even if the returns take years to materialise.
This story was first covered by Reuters.
