Elon Musk's xAI Activates World's Most Powerful AI Training System
The Colossus AI cluster, built with 100,000 Nvidia GPUs, aims to revolutionize AI capabilities and will soon double in size.
- xAI's Colossus cluster was brought online in just 122 days, setting a new benchmark in AI training systems.
- The system currently uses 100,000 Nvidia H100 GPUs and is expected to double to 200,000 GPUs soon.
- Colossus will train xAI's large language model, Grok, which competes with leading AI models like ChatGPT.
- Musk diverted GPU shipments from Tesla to xAI, potentially impacting Tesla's AI and robotics projects.
- The rapid development of Colossus has raised concerns about resource strains in Memphis, where the data center is located.