Meta Platforms is making a major move in the world of artificial intelligence. The company is putting the finishing touches on a new AI supercomputer, fueled by a staggering number of NVIDIA H100 AI GPUs—over 100,000 to be exact! This massive cluster, set to be operational by October or November, is rumored to be located somewhere in the United States.
The Information reports that this supercomputer will be the driving force behind training the next version of Meta’s Llama language model, aptly named Llama 4. This ambitious project is a testament to Meta’s commitment to advancing AI capabilities and pushing the boundaries of what’s possible.
The financial investment in this project is significant. The NVIDIA H100 AI GPUs alone are said to have cost Meta over $2 billion, showcasing the company’s willingness to pour resources into this AI venture. This massive investment speaks volumes about Meta’s belief in the power of AI and its potential to shape the future.
The collaboration between Meta and NVIDIA is undeniably strong. During a recent conversation, NVIDIA CEO Jensen Huang revealed that Meta now boasts a collection of over 600,000 NVIDIA H100 AI GPUs. Meta CEO Mark Zuckerberg responded with a chuckle, acknowledging that they are indeed “good customers” for NVIDIA. The sheer scale of this partnership underscores the importance of cutting-edge AI technology for Meta and its future endeavors.
