In the "Bg2Pod" podcast episode aired on October 13th, Nvidia CEO Jensen Huang remarked that setting up a supercomputer cluster in such a short time was "superhuman." Normally, this kind of work would take years.

"As far as I know, there's only one person in the world who could do that; Elon is uniquely knowledgeable about engineering, construction, large systems, and resource allocation; it's unbelievable," said the Nvidia CEO.

2crjfsej.png
Elon Musk was praised by Nvidia CEO Jensen Huang for setting up the supercomputer cluster in a short period of time. Photo: Bloomberg

xAI built the Colossus supercomputer from a cluster of 100,000 Nvidia H100 GPUs. In September, Elon Musk wrote in an X-blog post that it took 122 days “from start to finish” to bring the cluster online. He seemed to be referring to the total project time.

Nvidia's CEO also praised xAI's engineering, software, networking, and infrastructure teams as "extraordinary."

In an interview in June, Musk shared that it took Colossus 19 days to get from hardware installation to the start of training. He proudly stated that this was the fastest speed ever.

Furthermore, Huang also stated that xAI's supercomputer cluster is the fastest supercomputer on the planet. "A typical supercomputer takes three years to plan, ship equipment, and another year to become operational."

Musk launched the startup xAI in 2023 to compete with OpenAI. In August, the company introduced the AI ​​chatbot Grok-2. In early 2024, the company announced it had successfully raised $6 billion.

(According to Insider)