Elon Musk's xAI Unveils Most Powerful AI Training System

According to Cointelegraph, billionaire Elon Musk, owner of the social media platform X, announced that his AI development and research company, xAI, has brought its Colossus 100k H100 training cluster online. Musk revealed on Sept. 2 that it took his team 122 days to complete the process, and over the next few months, the system will double in size to 200k (50k H200s). Colossus is now considered the most powerful AI training system available, based on the number of graphics processing units (GPUs) powering it. One X user highlighted that this development surpasses every major model released to date, including OpenAI’s most powerful model, which uses 80,000 GPUs. The model was developed in collaboration with Nvidia, the leading semiconductor chip manufacturer. Nvidia’s H200, equipped with 141 GB of HBM3E memory and 4.8 TB/sec of bandwidth, was recently surpassed by the company’s latest Blackwell chip, unveiled in March 2024. Nvidia congratulated Musk and the xAI team, noting the system's exceptional gains in energy efficiency. The unveiling of Colossus has garnered significant attention across social platforms, with industry users calling it a major accomplishment in AI. Cathie Wood, CEO of ARK Invest, also praised the team, calling the achievement impressive and hinting at big announcements ahead. Cointelegraph has reached out to both Nvidia and xAI for comments on the model’s release. In April 2023, it was reported that Musk was purchasing large amounts of GPUs, with some sources indicating nearly 10,000 units, to further his xAI projects. Musk has been vocal about his plans for AI development, including a supercomputer that could be operational by fall 2025 and a potential partnership with Oracle. While leading the AI revolution, Musk has also supported regulations for safe AI. On Aug. 27, Musk posted on X advocating for California's Senate Bill 1047, a controversial stance in the tech industry due to concerns it could stifle AI innovation in the US. He has been an advocate for regulating AI for over two decades, comparing it to regulations on any product or technology that poses a potential risk to the public. Additionally, on Aug. 9, X agreed to suspend the processing of personal data for its users in the EU region following legal proceedings in an Irish court involving the local Data Protection Commission.