Elon Musk, a prominent American businessman, recently unveiled plans for his artificial intelligence startup xAI during a presentation to investors. According to The Information, Musk revealed that xAI is working on developing a supercomputer to power the next version of its AI chatbot Grok. He aims to have this cutting-edge technology operational by the fall of 2025 and is considering partnering with Oracle for its development.
The proposed supercomputer would consist of interconnected groups of Nvidia’s H100 GPUs, which are renowned for their exceptional performance in the data center chip market for AI. Musk announced in the presentation to investors that these GPU clusters would be at least four times larger than the largest ones currently available. However, obtaining these GPUs may pose challenges due to high demand.
xAI and Oracle did not provide immediate comments on the reported plans. Musk founded xAI last year as a competitor to Microsoft-backed OpenAI and Google’s Alphabet, in addition to being a co-founder of OpenAI himself. He mentioned earlier this year that training the Grok 2 model required 20,000 Nvidia H100 GPUs and that future iterations like Grok 3 would demand even more – around 100,000 Nvidia H100 chips.
Overall, Musk’s xAI is pushing boundaries in AI capabilities with its ambitious plans for a supercomputer to support its advanced chatbot technology.
