Nvidia’s Jensen Huang is all praises for Elon Musk’s xAI supercomputer: ‘Superhuman’ feat

The Colossus supercomputer is built from a cluster of 100,000 Nvidia GPUs. Even though Elon Musk had said that Colossus will take 19 days to build, the entire project was completed in 122 days from start to finish.

Jensen Huang praised xAI’s engineering, networking, infrastructure and software teams, referring to them as “extraordinary”. By putting 100,000 GPUs, Colossus has become “easily the fastest supercomputer on the planet as one cluster”, he said. 

It would normally take a lot more time for the supercomputer to be up and running, he said, adding, “A supercomputer that you would build would take normally three years to plan and then they deliver the equipment and it takes one year to get it all working.”

This comes as Nvidia’s shares roared back after the company successfully calmed investor concerns about product delays and its long-term growth prospects. The stock is up almost 14% this month and is the second-best performer in the S&P 500 Index this year.

Jensen Huang said Nvidia’s Blackwell chip “is in full production,” and that demand for it “is insane,” comments that came after Blackwell was delayed due to engineering snags, prompting a selloff that has now been erased.