Elon Musk has revealed ambitious plans to construct a “gigafactory of compute,” a term he coined to describe a vast supercomputer that would power an advanced version of xAI’s Grok. This initiative, as reported by The Information, intends to significantly enhance Grok’s capabilities, currently at version 1.5.
Musk’s vision involves assembling a colossal cluster of NVIDIA H100 GPUs, estimating a staggering 100,000 chips necessary for Grok 3.0. This figure marks a substantial increase, quintupling the 20,000 GPUs reportedly employed in the development of Grok 2.0. The projected cost of this endeavor runs into the billions, signaling Musk’s commitment to leading the frontier of AI technology.
Grok, which has evolved to process both visual data and text, has made notable advancements since its last update in April. As part of its ongoing development, xAI recently launched AI-generated news summaries for its premium users, showcasing Grok’s enhanced processing powers.
In discussions with investors, Musk highlighted that the planned GPU cluster would be at least four times larger than any current setup utilized by xAI’s industry rivals. This scale not only underscores the ambitious scope of Musk’s project but also his strategy to position Grok as a formidable force in the AI landscape.
Related News:
Featured Image courtesy of The Hollywood Reporter via Getty Images