Elon Musk's xAI built a 100,000-GPU supercluster in just 19 days - normally takes years
xAI's supercluster with 100,000 GPUs built in 19 days impressed Nvidia CEO.
Elon Musk's xAI team achieved a remarkable feat by constructing a supercluster consisting of 100,000 Nvidia H200 Blackwell GPUs in just 19 days. Normally, assembling such a computing powerhouse would take years, involving extensive planning, shipping, installation, and operational preparations.
Nvidia CEO Jensen Huang praised this record-breaking accomplishment as 'superhuman' and emphasized the complexity involved in integrating Nvidia's advanced hardware. The project required not only installing GPUs but also building a new 'X factory' with sophisticated cooling systems and power infrastructure to support the GPUs.
This rapid and precise execution positioned xAI as a formidable entity in AI research and development fields. The significant computational power unlocked could propel Musk's teams forward in developing cutting-edge neural networks, deep learning models, and natural language processing applications.