Wrong (it's 350K according to Zuck on insta) and dupe of "Meta will have more than 350.000 Nvidia H100s this year"[0], or "Zuckerberg's Meta Is Spending Billions to Buy 350k Nvidia H100 GPUs"[1] [0]: https://news.ycombinator.com/item?id=39044911 [1]: https://news.ycombinator.com/item?id=39046193
Meta buys 600k H100s to train LLaMa3
11–16 of 16 posts
Re: Meta buys 600k H100s to train LLaMa3
#12Re: Meta buys 600k H100s to train LLaMa3
#13Wrong (it's 350K according to Zuck on insta) and dupe of "Meta will have more than 350.000 Nvidia H100s this year"[0], or "Zuckerberg's Meta Is Spending Billions to Buy 350k Nvidia H100 GPUs"[1] [0]: https://news.ycombinator.com/item?id=39044911 [1]: https://news.ycombinator.com/item?id=39046193
350k in new H100s and in total 600k in H100s equivalents.
Re: Meta buys 600k H100s to train LLaMa3
#14I doubt they're paying retail prices like the tweet implies. Still can't be a small number if that is true.
Though as a wise man [1] once said: "The more you buy, the more you save".
[1] Jensen Huang, NVidia's CEO, says this quite often at conferences, its become a meme at this point, one example https://twitter.com/NVIDIAAI/status/917681161391841280
Re: Meta buys 600k H100s to train LLaMa3
#15https://en.wikipedia.org/wiki/List_of_countries_by_GDP_(nomi... lol, that's an investment!
Meta's annual revenues alone exceed $100B+ at this point, assuming these purchases are staggered over 2-3 years, it is basically adding a 2-3% overhead to their expenses on an annualized basis.
Re: Meta buys 600k H100s to train LLaMa3
#16Earlier quoted context omitted.
350k in new H100s and in total 600k in H100s equivalents.
I wonder what he means by "H100 equivalents" - are these GPUs built in-house or from competitors like AMD / Intel that give equivalent performance to Nvidia's H100s?
At the end of the day, the bottleneck is TSMC, on which all GPUs are produced right now, even the ones from Intel.