Mark Zuckerberg announced the construction of several multi-gigawatt datacenter clusters, starting with Prometheus set to debut in 2026. The clusters include Hyperion, which can scale to 5GW. The facilities will occupy significant land, comparable to a large portion of Manhattan. Zuckerberg aims to create a leading team for superintelligence development and plans to invest hundreds of billions in computing resources to support this initiative, promising high compute power per researcher in Meta Superintelligence Labs.
Meta is constructing several multi-gigawatt datacenter clusters, with the first named Prometheus set to launch in 2026, and plans for Hyperion to scale up to 5GW.
One of these datacenter clusters will cover a significant area, comparable to 80 percent of Manhattan's width, illustrating the massive scale of Meta's ambitions.
Zuckerberg emphasizes a focus on assembling a top-tier team for superintelligence development and investing hundreds of billions into computing resources to achieve these goals.
Collection
[
|
...
]