Elon Musk’s xAI has launched Colossus, now the most powerful AI training system, featuring a 100,000 H100 GPU cluster that will soon double to 200,000 GPUs. This marks a major step forward in the race to develop cutting-edge artificial intelligence. Musk announced on Sept. 2 that the xAI team took 122 days to bring Colossus online. The model is set to surpass the power of existing AI systems, including OpenAI’s, which runs on 80,000 GPUs.
Colossus was developed in collaboration with Nvidia, using the H200 semiconductor, known for its advanced performance and energy efficiency. Nvidia praised xAI’s achievement, highlighting that Colossus will be the most powerful AI system available.
Musk’s ambitions in AI development have been evident since early 2023, when reports surfaced about his massive GPU purchases to support xAI projects. His vision includes a supercomputer expected to be operational by 2025, in partnership with Oracle.
While leading the AI revolution, Musk has also called for responsible regulation. He recently expressed support for California’s Senate Bill 1047, which promotes AI safety, a stance that has sparked debate within the tech industry due to concerns about hindering innovation.
Colossus represents Musk’s latest move in the rapidly advancing AI landscape, setting a new benchmark for the field and positioning xAI as a key player in the future of artificial intelligence.