Jensen Huang, the CEO of NVIDIA recently appeared in an interview with the Tesla Owners Silicon Valley group, and one of the topics discussed was about Elon Musk's plan to develop a supercomputer cluster powered by 100 thousand NVIDIA H200 AI graphics cards.
He said that the will and ability to develop a supercomputer center with such a large number of graphics cards is not something new. However, what makes this unique is that Elon Musk and his team at X.AI and NVIDIA seem to have joined forces to do it all in just 19 days.
Jensen Huang said that normally, a data center with this kind of processing capacity would take around four years to develop. Three years will be required for the planning and design phase of this data center to ensure that power and cooling are sufficient to power this data center, and so that it can be used with the company's existing software and technology. Another year is usually required to build the infrastructure.
However, Elon Musk is not a patient person, and this supercomputer cluster infrastructure was discussed, designed and built all in just about two weeks. Jensen Huang is also confident that during those 19 days, all individuals involved in the development of this data center did not have time to sleep, but to ensure that X.AI can operate as quickly as possible, this is what Elon Musk did.
What is interesting about this rapid development is how X.AI will leverage their technology for the purpose of developing their own artificial intelligence technology. As Jensen himself said, the process of building infrastructure is usually not something that can be done in a short amount of time.
There are many considerations that need to be made to ensure that everything works properly. We may soon be able to see for ourselves how the company will harness the power of the world's most powerful supercomputer cluster.