Mark Zuckerberg Reveals Major Meta AI Supercluster Push
In a sweeping announcement on Facebook, Meta CEO Mark Zuckerberg outlines plans to sink “hundreds of billions of dollars” into building artificial superintelligence—AI systems capable of outperforming human intelligence.
And data centres are at the heart of that ambition.
Meta is already rolling out two groundbreaking projects—Prometheus and Hyperion—designed to deliver industrial-scale compute through superclusters.
These data centres will be the physical backbone for Meta’s AI push, providing the bandwidth, processing power and technical architecture to train its next-generation models.
“Meta Superintelligence Labs will have industry-leading levels of compute and by far the greatest compute per researcher,” says Mark.
His vision depends not just on hiring elite researchers, but on building facilities that can match their ambitions with compute power and scale.
Building the foundations of superintelligence
The centrepiece of this strategy is Prometheus, a 1 gigawatt (GW) data cluster set to go online in 2026. That figure alone places it ahead of any existing AI infrastructure. According to SemiAnalysis, Meta will become the first lab to launch a 1 GW+.
At an estimated US$30bn per GW, Meta’s compute ambitions are groundbreaking.
Hyperion follows as a longer-term project, designed to be scalable up to 5 GW across multiple phases spanning several years.
Both superclusters aim to support the demands of Meta’s expanding model suite, including upcoming versions of Llama 4 and beyond.
Alongside Prometheus and Hyperion, Meta is also preparing a series of “titan” clusters to further increase total capacity.
Meta’s pivot to data infrastructure echoes rivals like Microsoft and Amazon, which are already partnering with major AI developers while massively scaling their own compute.
But in his announcement, Mark insists Meta’s scale and independence give it a unique edge, stating: “We have the capital from our business to do this.”
Talent, tools and technical firepower
To populate these clusters with purpose, Meta has launched Meta Superintelligence Labs (MSL), a new division uniting its AI efforts under the leadership of Alexandr Wang, formerly of Scale AI, Daniel Gross, the CEO and co-founder of Safe Superintelligence and Nat Friedman, ex-GitHub CEO.
The division already benefits from Meta’s US$14.3bn acquisition of Scale AI, known for its data labelling infrastructure and the HLE (Humanity’s Last Exam) benchmark. This acquisition also addresses quality concerns that troubled Llama 4.
Meta has been embarking on a multi-billion-dollar recruitment campaign to hire talent from leading tech companies to push ahead.
New recruits include talent from Google DeepMind, OpenAI and Anthropic, including Hongyu Ren (co-creator of GPT-4o and o-series models), Jiahui Yu (GPT-4o co-creator), Jack Rae (Gemini pre-training lead), Pei Sun (Gemini reasoning expert) and Shengjia Zhao (ChatGPT co-creator).
Meta has reportedly offered packages as high as US$300m over four years, sparking disputes across the industry.
Zuckerberg’s personal involvement—meeting candidates at his Lake Tahoe and Palo Alto homes—signals how MSL is central to Meta’s identity going forward.
Previously, in an internal memo to Meta staff, Mark said: “As the pace of AI progress accelerates, developing superintelligence is coming into sight. I believe this will be the beginning of a new era for humanity and I am fully committed to doing what it takes for Meta to lead the way.”
In this recent announcement, Mark adds clarity on how he envisions the company making this a reality: “For our superintelligence effort, I'm focused on building the most elite and talent-dense team in the industry.”
In response to the announcement, Karl Freund, Principal Analyst at Cambrian AI Research, told the BBC, "Clearly, Zuckerberg intends to spend his way to the top of the AI heap."
"The talent he is hiring will have access to some of the best AI Hardware in the world," he added.
