Elon Musk and the team behind xAI have achieved an engineering marvel, setting up a supercluster of 100,000 H200 Blackwell GPUs in a whopping 19 days. Nvidia CEO Jensen Huang told the story of Elon Musk's incredible installation prowess with members of the Tesla Owners Silicon Valley on X.
Huang describes Musk's 19-day escapade with awe and respect, calling the effort "superhuman". The team at xAI purportedly went from the "concept" phase to full-ready compatibility with Nvidia's "gear" in less than three weeks. This includes running xAI's first AI training run on the newly built supercluster as well.
Elon Musk is super human. What would take everyone else a year, only took him 19 days. pic.twitter.com/q51sM48lsuOctober 13, 2024
From start to finish, the process involved building the massive X factory where the GPUs would reside and equipping the entire factory with liquid cooling and power to make all 200,000 GPUs operational. That's not to mention all of the coordination between Nvidia's and Elon Musk's engineering teams to get all of the hardware and infrastructure shipped and installed precisely and in a coordinated manner.
For perspective, Huang states that it takes an average data center four years to do what Elon Musk and his team were able to do in 19 days. Three years of that time alone would be dedicated to planning, while the last year would be used to ship the equipment, install it, and get it all working.
Huang also goes into detail describing how complex the networking is on Nvidia's hardware. He explains that networking Nvidia's gear isn't like networking traditional data center servers. "The number of wires that goes in one node...the back of a computer is all wires."
Elon Musk's integration of 100,000 H200 GPUs has "never been done before" (according to Jensen Huang) and probably won't be duplicated again by another company, at least not for a very long time.
Stay On the Cutting Edge: Get the Tom's Hardware Newsletter
Get Tom's Hardware's best news and in-depth reviews, straight to your inbox.
Aaron Klotz
Contributing Writer
Aaron Klotz is a contributing writer for Tom’s Hardware, covering news related to computer hardware such as CPUs, and graphics cards.
More about gpus
Latest
56 CommentsComment from the forums
newtechldtech Elon Mask what ? he did not do anything but hire people and pay $$$ .
Reply
great Unknown newtechldtech said:
Elon Mask what ? he did not do anything but hire people and pay $$$ .Jensen Huang apparently disagrees with you. But what does Jensen know? All he does is hire people and pay $$$.
And apparently hiring people and paying $$$ is not as simple as it sounds, or other corporations would have done the same.
Reply
JTWrenn great Unknown said:
Jensen Huang apparently disagrees with you. But what does Jensen know? All he does is hire people and pay $$$.And apparently hiring people and paying $$$ is not as simple as it sounds, or other corporations would have done the same.
It's hype. That is how ceo's work. Push employees like mad and cut corners then take credit for their work.
I will wait to see how screwed up it is for that crunch rather than believe two ceo's patting each other on the back.
Reply
padrescout great Unknown said:
Jensen Huang apparently disagrees with you. But what does Jensen know? All he does is hire people and pay $$$.And apparently hiring people and paying $$$ is not as simple as it sounds, or other corporations would have done the same.
Or there is a reason it's not done this way. Wouldn't be the first time Musk decided he knows better, ignores the institutional knowledge a industry had built and then has to make exactly all those mistakes until he ...does it the way everyone else does. See ... everything Tesla has ever done in vehicle manufacturing. Or Musk demanding they stop a bunch of "useless services" and breaking Twitter... repeatedly. Or his bad code he writes that is just .... funny. That ones mostly funny, no-one uses his code. He's a terrible coder by all accounts. "Print your code and come explain it to me!" That's just funny. He's a laughable person. A sad laughable half-smart person.
So yeah, let's give it 6 months and see what unforeseen consequences, if any, reveal themselves.
Reply
acadia11 Admin said:
Elon Musk and the team behind xAI purportedly setup a total of 100,000 H200 Nvidia GPUs in just 19 days. That's a feat that should have taken four years to complete.Elon Musk took 19 days to set up 100,000 Nvidia H200 GPUs; process normally takes 4 years : Read more
No, the vision and drive to make that happen takes a pretty big effort, and I’m quite confident if you were given the money and the hires you still couldn’t get it done. If everyone could do it they would.
I would say Huang and his teams also deserves a great deal of credit.
Reply
domih What would Elon Musk do? /s
Reply
ingtar33 newtechldtech said:
Elon Mask what ? he did not do anything but hire people and pay $$$ .you know i keep seeing people say this.
tesla does something amazing "elon didn't do it, he hired people to do it"
spacex does a once in a lifetime engineering marvel "elon didn't do it, he hired people to do it"
elon sets up an ai datacenter in 19days that normally takes 4 years "elon didn't do it, he hired people to do it"Even if you're right and elon is not a competent engineer (something i think his career disproves but whatever, i won't argue it), he certainly might be the most competent hiring manager in the history of mankind. you only get this lucky once in your life and he keeps doing it.
Reply
DavidLejdar Wouldn't want to live nearby:
https://www.reuters.com/business/environment/musks-xai-operating-gas-turbines-without-permits-data-center-environmental-group-2024-08-28/
Btw, I can set up a chat-bot, which shouts: "1984!", in under an hour. Just sayin'. :D
(This is not a political reference - I don't even live in the U.S. It is just that the AI model is named after The Hitchhiker's Guide to the Galaxy, novels I like, and I also like Dirk Gently’s Holistic Detective Agency, and the videogame Starship Titanic - and then I look at Mr Musk's... uhm... political activism, and it just happens to remind me rather of a different novel. Which I just meant to express, that some of that stuff in the U.S. seems quite absurd to me.)ingtar33 said:
...Even if you're right and elon is not a competent engineer (something i think his career disproves but whatever, i won't argue it), he certainly might be the most competent hiring manager in the history of mankind. you only get this lucky once in your life and he keeps doing it.
Huh? E.g. Bosch (named after a guy) has a lot more employees than i.e. Tesla. Likewise, e.g. Siemens, also named after a guy, with e.g. the first electric locomotive having been here in Berlin (and then having ended up as something, not just a selected few had some use of).
Not meaning to argue about whether Mr Musk really is the best the U.S. has. But "in the history of mankind", there are a lot more examples, than a library visitor in Florida may be allowed to learn about.
Reply
why_wolf More like most companies will spend 3 years budgeting for a project and/or buying parts for it on a piecemeal schedule and then go live at the end.
Most companies don't just randomly spend a few billion dollars out of nowhere and pay the price premium for the rush job because the CEO is riding the dragon again.
Reply
Stomx And what happened with the development of Musk own chips? Instead he spent $4B in 19 days for 10x-20x overpriced chips.
Reply
