Meta’s massive AI bet: Zuckerberg unveils superclusters to build artificial superintelligence; key takeaways here

Meta is ramping up its AI ambitions with plans to invest hundreds of billions of dollars to build several multi-gigawatt data centre clusters aimed at developing artificial superintelligence. CEO Mark Zuckerberg revealed the scale of the spending on Monday, stating that the investment would be backed by the $100 billion-a-year ad-driven giant’s robust advertising business.

“For our superintelligence effort, I'm focused on building the most elite and talent-dense team in the industry. We're also going to invest hundreds of billions of dollars into compute to build superintelligence. We have the capital from our business to do this,” Zuckerberg said in a Facebook post late Monday evening.

For now, the tech giant is involved in building multi-gigawatt compute clusters. Zuckerberg confirmed that the first one, called Prometheus—which means “forethought” in Greek—will be built by 2026.

“We're actually building several multi-GW clusters. We're calling the first one Prometheus, and it's coming online in '26. We're also building Hyperion, which will be able to scale up to 5 GW over several years. We're building multiple more titan clusters as well. Just one of these covers a significant part of the footprint of Manhattan,” Zuckerberg said in the post.

Additional clusters are also in development as part of a broader infrastructure push to support large-scale AI model training.

A multi-gigawatt supercluster refers to a massive data centre complex designed to train and run advanced artificial intelligence models at an unprecedented scale.

These data centres will support the work of Meta’s newly formed Superintelligence Labs division, created last month following internal AI setbacks, including the underwhelming release of Llama 4 and the departure of key engineers. The reorganisation reflects Zuckerberg’s renewed focus on superintelligence and his dissatisfaction with Meta’s previous AI trajectory.

As reported by Fortune India earlier, Superintelligence Labs will be led by Alexandr Wang, former CEO of Scale AI, and Nat Friedman, former GitHub chief. Meta recently invested $14.3 billion in Scale AI and has launched an aggressive hiring spree to build what it calls the most elite AI research team in the industry. The company has also offered to purchase a minority stake in funds run by Friedman and Daniel Gross via a tender offer to existing investors.

The newly formed Superintelligence Labs is expected to generate returns through multiple products, including the Meta AI assistant, ad tools that convert images into videos, and AI-powered smart glasses.

Zuckerberg emphasised that Meta will soon have the highest compute capacity per researcher in the industry, underscoring its ambitions in building general-purpose AI systems that surpass human-level intelligence.

According to a report by chip industry publication SemiAnalysis, Meta is on track to become the first AI lab to bring a 1-gigawatt-plus supercluster online. Zuckerberg also confirmed the news in his Monday post. The SemiAnalysis report further reveals that Hyperion, Meta's Louisiana facility, is expected to become the world’s largest individual AI data centre campus by the end of 2027, with over 1.5 gigawatts of IT power in its first phase—surpassing OpenAI’s Stargate capacity.