Meta's CEO, Mark Zuckerberg, announced plans to invest hundreds of billions of dollars in building several large artificial intelligence (AI) data centers to develop what he describes as "superintelligence." The social media giant is actively recruiting AI talent by offering lucrative salaries and striking high-profile deals to quickly develop machines potentially smarter than humans in various tasks.
According to Zuckerberg's post on his social media platform, Threads, Meta's first multi-gigawatt data center, named "Prometheus," is expected to go online in 2026. Another data center, Hyperion, is projected to expand to 5 gigawatts in the coming years. Additionally, Meta is constructing more Titan clusters, one of which will cover a significant portion of Manhattan.
Despite investor concerns over the return on these massive expenditures, Zuckerberg justified the spending by highlighting the strength of Meta's core advertising business. The company recently restructured its AI division, forming a new entity called Superintelligence Labs, following setbacks with its open-source model Llama 4 and key personnel departures.
Meta is betting on generating new revenue streams from its AI applications, image-to-video ad tools, and smart glasses. Reports suggest that the company is considering replacing its most powerful open-source AI model, Behemoth, with a closed alternative.
Analyst Gil Luria from D.A. Davidson noted that Meta's significant AI investments are driven by the technology's impact on its advertising business, allowing for more lucrative ad sales. However, achieving a leading AI model may take time. Recently, Meta invested $14.3 billion in Scale AI and launched a talent acquisition campaign for its Superintelligence Labs, led by former executives from Scale AI and GitHub.
