Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Hetzner introduces GPU server for AI training

hetzner.com

Post reply on HN
Print view View original

21–30 of 57 posts

Re: Hetzner introduces GPU server for AI training

#21

Post by rafaelmn »

jsheard wrote: post #17
andersa wrote: post #12

Earlier quoted context omitted.

Not before the bubble pops.

Even when it does, GPUs with more VRAM than a flagship gaming model have always and will always come at a massive price premium. That ceiling is currently 24GB, if you need more than that it's going to cost you.

I don't see why. AMD could buy a lot of free community work just by putting out a 32 GB version of their next gen flagship at around 1k. Not chasing the margin for a generation would buy them a decent amount of community support - they need it if they want to compete with Nvidia at any point.

Ahh I just saw OP was saying sub 100$ - yeah that's never going to happen.

Re: Hetzner introduces GPU server for AI training

#24

Post by kirab »

andersa wrote: post #6

Hmm. Seems like a bad deal. This is a monthly reservation for a single 6000 Ada for $940. You can get the same on RunPod for $670. And to actually train stuff you'd likely want nodes with more of them, like 8, or just different GPUs all together (like A100/H100/etc).

Hetzner customer here. It’s a little hard to understand in the UX, but the price shown is the monthly max price.

It is actually paid by the hour.

The price per hour for this server is € 1.5980

more info: https://docs.hetzner.com/general/others/new-billing-model/

Re: Hetzner introduces GPU server for AI training

#25

Post by kirab »

mromanuk wrote: post #22

I use runpod or vast for training my (small) models (a few million parameters) mostly using RTX4090 up to 4 GPUs. Training is a sporadic task. Is not worth it for me to book it monthly (at these prices)

Hetzner customer here. It’s a little hard to understand in the UX, but the price shown is the monthly max price.

It is actually paid by the hour.

The price per hour for this server is € 1.5980

more info: https://docs.hetzner.com/general/others/new-billing-model/

Re: Hetzner introduces GPU server for AI training

#26

Post by InvaderFizz »

rafaelmn wrote: post #21
jsheard wrote: post #17

Earlier quoted context omitted.

Even when it does, GPUs with more VRAM than a flagship gaming model have always and will always come at a massive price premium. That ceiling is currently 24GB, if you need more than that it's going to cost you.

I don't see why. AMD could buy a lot of free community work just by putting out a 32 GB version of their next gen flagship at around 1k. Not chasing the margin for a generation would buy them a decent amount of community support - they need it if they want to compete with Nvidia at any point. Ahh I just saw OP was saying sub 100$ - yeah that's never going to happen.

VRAM is such a small portion of the BOM that they could double it, keep the same margins, and the MSRP would only increase by like $300.

The problem isn't margin on the consumer cards, the problem is competing with their even higher margin workstation cards.

Re: Hetzner introduces GPU server for AI training

#27

Post by naiv »

kirab wrote: post #25
mromanuk wrote: post #22

I use runpod or vast for training my (small) models (a few million parameters) mostly using RTX4090 up to 4 GPUs. Training is a sporadic task. Is not worth it for me to book it monthly (at these prices)

Hetzner customer here. It’s a little hard to understand in the UX, but the price shown is the monthly max price. It is actually paid by the hour. The price per hour for this server is € 1.5980 more info: https://docs.hetzner.com/general/others/new-billing-model/

But you have to pay the setup fee of 94 Euros every time you would choose this server, or?

Re: Hetzner introduces GPU server for AI training

#28

Post by reitzensteinm »

kirab wrote: post #24
andersa wrote: post #6

Hmm. Seems like a bad deal. This is a monthly reservation for a single 6000 Ada for $940. You can get the same on RunPod for $670. And to actually train stuff you'd likely want nodes with more of them, like 8, or just different GPUs all together (like A100/H100/etc).

Hetzner customer here. It’s a little hard to understand in the UX, but the price shown is the monthly max price. It is actually paid by the hour. The price per hour for this server is € 1.5980 more info: https://docs.hetzner.com/general/others/new-billing-model/

But the setup fee is €86.90. Is that not paid each time you provision the server?

Re: Hetzner introduces GPU server for AI training

#29

Post by dindresto »

kirab wrote: post #24
andersa wrote: post #6

Hmm. Seems like a bad deal. This is a monthly reservation for a single 6000 Ada for $940. You can get the same on RunPod for $670. And to actually train stuff you'd likely want nodes with more of them, like 8, or just different GPUs all together (like A100/H100/etc).

Hetzner customer here. It’s a little hard to understand in the UX, but the price shown is the monthly max price. It is actually paid by the hour. The price per hour for this server is € 1.5980 more info: https://docs.hetzner.com/general/others/new-billing-model/

That's incorrect, dedicated Hetzner servers do not have hourly pricing. That only applies to Hetzner Cloud.

Re: Hetzner introduces GPU server for AI training

#30

Post by reitzensteinm »

dindresto wrote: post #29
kirab wrote: post #24

Earlier quoted context omitted.

Hetzner customer here. It’s a little hard to understand in the UX, but the price shown is the monthly max price. It is actually paid by the hour. The price per hour for this server is € 1.5980 more info: https://docs.hetzner.com/general/others/new-billing-model/

That's incorrect, dedicated Hetzner servers do not have hourly pricing. That only applies to Hetzner Cloud.

That's incorrect. Dedicated servers do now offer hourly pricing. See this page which lists the hourly pricing for the new GEX130 server:

https://www.hetzner.com/dedicated-rootserver/matrix-gpu/

But, as far as I know (hence the sibling question), you still need to pay the setup fee each time you launch.

Post reply on HN