Mark Zuckerberg announced Meta’s ambitious plan to invest “hundreds of billions of dollars” in constructing large AI data centers. This initiative aligns with Meta’s focus on achieving superintelligence—creating machines that surpass human capabilities in various tasks. In a bid to attract top talent, Meta is reportedly offering lucrative compensation packages to AI engineers from competitors like Google, Apple, and OpenAI. Zuckerberg emphasized that the company’s robust advertising revenue will fund these ventures, addressing investor concerns about the significant financial commitment. A report by SemiAnalysis indicated that Meta aims to be the first to launch a 1-gigawatt-plus supercluster for training advanced AI models. The initial data center, named Prometheus, is expected to be operational in 2026, while another, Hyperion, will have the capacity to scale to 5 gigawatts. Meta’s development of these multi-gigawatt clusters aims to support its vision of breakthrough advancements in AI technology.
Source link
