
Elon Musk, the renowned U.S. entrepreneur, revealed to investors through a presentation reported by The Information that his artificial intelligence venture, xAI, is gearing up to construct a supercomputer to enhance its AI chatbot, Grok. Musk expressed a targeted launch window for this advanced computing system by the fall of 2025. Additionally, he hinted at a potential collaboration with tech giant Oracle for the development of this ambitious project. Despite attempts to reach out, xAI remained unavailable for immediate comment, and Oracle refrained from responding to Reuters' inquiry.
According to Musk's presentation to investors in May, the envisioned supercomputer would leverage Nvidia's H100 GPUs, which are renowned for their dominance in the AI-focused data center chip market. Musk stated that the proposed clusters of H100 GPUs would surpass the scale of existing GPU clusters by at least fourfold. However, acquiring Nvidia's H100 GPUs can pose challenges due to their high demand in the market.
xAI, established by Musk last year, aims to rival established players such as Microsoft-backed OpenAI and Google's Alphabet. Notably, Musk has a history with OpenAI, having co-founded the organization. Earlier this year, Musk disclosed that training the Grok 2 model necessitated approximately 20,000 Nvidia H100 GPUs. He further projected that future iterations, such as Grok 3 and beyond, would demand an even more staggering quantity of 100,000 Nvidia H100 chips.
Comentarios