The founder of Facebook (owned by Meta), Mark Zuckerberg, has recently revealed his company’s intention of building large data centers to train AI. This project will potentially cost hundreds of billions of dollars and would allow Zuckerberg to realize the goal of achieving superintelligence.
See Related: Facebook Has A New Virtual Assistant: Meet Meta AI
As reported by Reuters, the parent company of Facebook is currently working on two separate data centers. The first one, designated as Prometheus, will be operational in 2026. The second of the two, Hyperion, will take a few years to be online but is estimated to scale up to 5 gigawatts.
“We’re building multiple more titan clusters as well. Just one of these covers a significant part of the footprint of Manhattan”, Zuckerberg posted on his account on Thread. While the cost of the whole project has raised eyebrows, the billionaire CEO said that he and his company have the capital to fund it.
After several setbacks to its AI ambitions, Meta has refocused its efforts on artificial intelligence in the past months. It has established the Superintelligence Lab, a division devoted to pushing Meta’s AI capabilities to the top of the industry as it competes with other tech companies for AI supremacy. To facilitate this new endeavor, Meta is hoping that it can generate sufficient cash flow from its other products, such as the Meta AI app, smart glasses, and more.