Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


Shells Virtual Desktop
BMail.ag - Secure Email Service
Server.net
CPLicense.net
VPS Server
Buy VPN
Vultr
VMs for AI
HostDare
ReliableSite White-Label Dedicated Hosting for Resellers
InterServer VPS
BMail.ag - Secure Email Service
Best VPN
High-Performance Bare Metal Server Solutions
Karvl.com
Server Mania Cloud Hosting
DataWagon Hosting
AlphaVPS Hosting
Evoxt.com
Clouvider
VPS Hosting with NVMe
Residential IPs in the US & 4G Mobile Proxies in EU & US with Unlimited Bandwidth
ReliableSite White-Label Dedicated Hosting for Resellers
Rabisu - Hosting Solutions
Shells Virtual Desktop

In this Discussion

New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

Will server hardware get really cheap in a few years if some AI companies go under?

inkkoinkko Member

A few days ago, BreezeHost launched another FLASH SALE for their Mystery Box dedicated servers, with a wide range of datacenter locations to choose from. It makes me wonder if some company recently went bankrupt and sold off all their dedicated servers in bulk to BreezeHost, which is why they’re able to run this kind of Mystery Box promotion again.

From what I’ve seen, most people didn’t get particularly great specs, but I did come across someone who scored a dual Gold CPU + 192GB RAM + 8TB NVMe. Honestly, I’m pretty jealous—but I don’t really want to gamble on it. (That said, if OVH does Mystery Box for Easter next month, I’ll probably still give it a shot.)

I also can’t help but wonder—will DDR5 dedicated servers become significantly cheaper in a few years? Or will we see even more Mystery Box flash sales?

Thanked by 1oloke

Comments

  • ServerAstraServerAstra Member, Patron Provider

    I don't believe AI companies will ever go bankrupt as mostly they are backed by NASDAQ, DOW and S&P members which are very stable.
    Same as during Bitcoin craze somehow all the videocards were bought out and suddenly we have quite smart AI now, but still not much bulk videocards on the market compared to the shortage, meaning they were sent somewhere else and now are still used.
    I might sound like a conspiracy theorist but I'm quite sure all of this is more related to Anduril and other overlords' endeavours similar to that (ie newer global positioning system which requires a lot of ram to keep the whole magnetic map of earth in the memory), just done via "special" channels. There was no profit at all in OpenAI grabbing 40% of consumer and server DDR5 memory market production (compared to GPUs it is extremely slow), and pretty sure a company on the brink and basically being "donation-ware" cannot secure such funds anyhow unless there's some very high level support.

  • rpqurpqu Member

    @ServerAstra said:
    I don't believe AI companies will ever go bankrupt as mostly they are backed by NASDAQ, DOW and S&P members which are very stable.

    AI companies rent the cards

    Same as during Bitcoin craze somehow all the videocards were bought out and suddenly we have quite smart AI now, but still not much bulk videocards on the market compared to the shortage, meaning they were sent somewhere else and now are still used.

    Faulty, underclocked cards did flood the market. The rest aren't sold because it doesn't survive the mining environment.

  • ServerAstraServerAstra Member, Patron Provider

    @rpqu said:
    AI companies rent the cards

    Yep, I'm not arguing with that.

    Faulty, underclocked cards did flood the market. The rest aren't sold because it doesn't survive the mining environment.

    2012 (AlexNet Milestone): The breakout moment for deep learning was the realization that GPUs could train neural networks much faster than CPUs. AlexNet used two NVIDIA GTX 580 GPUs. This sparked an industry-wide realization that GPUs were essential for AI research.
    Mid-2010s (AI Research Boom): As researchers rushed to train larger models, GPUs became the indispensable tool, fueling high-end GPU sales.
    2016-2017 (Generative AI/LLM beginnings): Development of models requiring massive parallel processing continued to accelerate, making top-tier gaming cards (like the GTX 1080 Ti) highly sought after for both gaming and AI development.

    Cue the 2017-2018 Crypto Spike which coincided with Gaming spike and release of 1080 chips - which are still not flooding the market. Why? They are useless for AI. Now pick any card which is good for AI? Won't be able to find it (prominent non-speculative crypto don't care about VRAM anymore and no one is buying 5000 gpus to thrash Vertcoin). But used AI munchers? Sure, plenty now, though very old and power hungry compared to current ones. 5090 was difficult to get on release not because of Crypto. 4090 same, and 3090 most likely too.

Sign In or Register to comment.